For this blog post, I will explore some of the unsupervised learning methods from recent weeks (I worked on supervised learning methods for a bit but then realized it wasn’t the best strategy for my data and what I am trying to do). At this point, I plan to make the next (and last) blog post 6 a stepping stone to my final project by examining what I have done so far and revisiting my research questions to review and make sure I have answered them.
Data
The dataset contains 457 different articles from Arlnow (local news site in Northern Virginia) from March 2020 to September 2022 (550 before removing “Morning Notes”). The code below comes from other blog posts so I will not explain it much, but it involves cleaning and pre-processing and generally getting the data ready to use.
said new cases covid covid-19 also school public
982 689 583 517 472 463 444 441
health year people community one can last local
435 410 391 391 385 382 376 342
week says pandemic two
335 333 328 310
Analysis - Unsupervised Learning (topic modeling)
Topic models are used to identify the topics that characterize a set of documents. LDA and STM are both mixed-membership models, meaning documents are characterized as arising from a distribution over topics, rather than coming from a single topic.
Latent Dirichlet Allocation
LDA uses two basic principles: each document is made up of topics, and each word in a document can be attributed to a topic.
Code
# create new LDA modellda_model <- LDA$new(n_topics =10, doc_topic_prior =0.1,topic_word_prior =0.01)# print other methods for LDAlda_model
<WarpLDA>
Inherits from: <LDA>
Public:
clone: function (deep = FALSE)
components: active binding
fit_transform: function (x, n_iter = 1000, convergence_tol = 0.001, n_check_convergence = 10,
get_top_words: function (n = 10, topic_number = 1L:private$n_topics, lambda = 1)
initialize: function (n_topics = 10L, doc_topic_prior = 50/n_topics, topic_word_prior = 1/n_topics,
plot: function (lambda.step = 0.1, reorder.topics = FALSE, doc_len = private$doc_len,
topic_word_distribution: active binding
transform: function (x, n_iter = 1000, convergence_tol = 0.001, n_check_convergence = 10,
Private:
calc_pseudo_loglikelihood: function (ptr = private$ptr)
check_convert_input: function (x)
components_: NULL
doc_len: NULL
doc_topic_distribution: function ()
doc_topic_distribution_with_prior: function ()
doc_topic_matrix: NULL
doc_topic_prior: 0.1
fit_transform_internal: function (model_ptr, n_iter, convergence_tol, n_check_convergence,
get_c_all: function ()
get_c_all_local: function ()
get_doc_topic_matrix: function (prt, nr)
get_topic_word_count: function ()
init_model_dtm: function (x, ptr = private$ptr)
internal_matrix_formats: list
is_initialized: FALSE
n_iter_inference: 10
n_topics: 10
ptr: NULL
reset_c_local: function ()
run_iter_doc: function (update_topics = TRUE, ptr = private$ptr)
run_iter_word: function (update_topics = TRUE, ptr = private$ptr)
seeds: 1550276485.55619 698001923.849935
set_c_all: function (x)
set_internal_matrix_formats: function (sparse = NULL, dense = NULL)
topic_word_distribution_with_prior: function ()
topic_word_prior: 0.01
transform_internal: function (x, n_iter = 1000, convergence_tol = 0.001, n_check_convergence = 10,
vocabulary: NULL
INFO [12:22:10.326] early stopping at 100 iteration
INFO [12:22:12.287] early stopping at 50 iteration
Code
# doc_topic_distr is a matrix where each row is a document, each row is a topic, and the cell entry is the proportion of the document estimated to be of the topic - each row is the topic attention distribution for a document# topic distribution for the first documentbarplot(doc_topic_distr[1, ], xlab ="topic",ylab ="proportion", ylim =c(0,1),names.arg =1:ncol(doc_topic_distr))
Code
# get top n words for all topicslda_model$get_top_words(n =5, lambda =0.2)
# quite a few of these topics actually seem to make sense and fit together - I think this was fairly successful and creates some interesting ideas and topics## visualization (opens in browser)#lda_model$plot()
Structural Topic Models
STM allows for leveraging the information (metadata) as part of the estimation of the topics. In this case, estimating topical prevalence by incorporating time/date information in estimating the topics.
Code
# correlated topic model - a structural topic model that incorporates covariates; an STM without covariates reduces to a very fast implementation of correlated topic models (a version of the vanilla LDA model but where the topic proportions can be positively correlated with one another).# creating a new variable from time_tag that will just be the year, then recreating the dfmtime_year =as.factor(substr(arlnow_covid$time_tag, 1, 4))arlnow_covid$time_year = time_yeararlnow_covid = arlnow_covid[-c(3)]is.character(arlnow_covid$time_year)
# top document associated with each topicfindThoughts(cor_topic_model,texts = arlnow_covid$text_field,topics =c(1:7),n =1)
Topic 1:
Exactly one month ago, the average Covid test positivity rate in Arlington was 3.5%. Today that rate is 26.1%., The positivity rate has been soaring amid a Covid wave fueled by the highly-contagious Omicron variant. The wave has produced sky-high case totals, but has yet to correspond to a surge in serious illness., Yesterday Arlington reported its highest seven-day moving average of new cases, 491 cases per day, before dropping today as the New Year’s holiday and the winter storm tamp down on testing. On Friday, before the holiday, Arlington set a fresh single-day case record, with 784 new Covid cases reported., The average daily rate of hospitalizations for Covid among Arlington’s highly-vaccinated population has risen slightly, to just over one per day, but that’s a far cry from the levels of serious illness earlier in the pandemic., No new deaths have been reported over the past week, per Virginia Dept. of Health data., While the hospitalization numbers in Arlington are relatively low, that’s not necessarily the case statewide. Hospitals across the Commonwealth “are becoming overwhelmed” amid rising illness, the Virginia Dept. of Health said last week., “The VDH urges everyone to reserve hospitals for emergencies,” the state health department said. “If you have mild coronavirus symptoms or a non-serious illness, avoid unnecessary hospital trips.”, Virginia Hospital Center ER chief Mike Silverman wrote in his weekly update this past Friday that the hospital is also seeing an increase in cases., Although Omicron appears less dangerous than Delta, last week our hospital inpatient COVID volume was 40% of our prior peak number. This week, it’s over 70%. Although the overall morbidity and mortality rates may be lower with Omicron across the population, the actual numbers of patients getting sick with COVID may exceed any hospital’s capacity to care for those requiring it. In the ER, we’ve set records when it comes to the numbers of patients we’ve diagnosed with COVID. COVID isolation orders have doubled in the past week and represents about half of our daily volume. Fortunately, the admission rate on these patients is about 15%. In previous surges, this number was 40%., Consistent with all the other places you’ve seen case counts, we’ve had a 10-fold increase in the number of positive covid tests in our ER compared to 2 weeks ago. About 41% of our patients are testing positive -this is about a 60% positivity rate when looking at symptomatic patients (was 13% one month ago) and 25% positivity when screening patients (admissions, transfers, non-covid symptoms). This number was <1% a month ago. The number of tests we perform weekly has doubled from a month ago while the positive numbers have increased 30-fold (this was tricky math for me, but I think it’s right)., Despite the continued rise in cases, the high positivity rate and challenges with testing suggest that there are potentially a lot more people getting infected than are getting tested and submitted to the VDH database as positive cases., After closures over the weekend for the New Year’s holiday, for instance, county Covid testing booths — which have seen huge lines for the past couple of weeks — are closed again today due to the winter storm., https://twitter.com/ReadyArlington/status/1478372102692560897?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm%5E1478372102692560897%7Ctwgr%5E%7Ctwcon%5Es1_&ref_url=https%3A%2F%2Fwww.arlnow.com%2F, At-home tests, meanwhile, remain hard to come by. At a CVS on Columbia Pike today, a sign says that due to the winter storm the store will not receive more test kits tomorrow, as originally expected., Read More
Topic 2:
, Ceremonies and events will be held in Arlington this weekend to commemorate the 20th anniversary of the 9/11 terrorist attacks on Saturday., The events will pay tribute to those who died as well as to Arlington’s first responders, whose response to the Pentagon attack has been hailed as a “model for the nation” by the 9/11 Commission., Among the events on tap are a wreath-laying ceremony, a memorial 5K, a bike ride, a softball tournament, and a private event at the Pentagon for the families of the victims., With some events at-capacity or closed to the public, the county says locals can observe the day from home, by watching short videos produced by the county, or by hanging American flags., Wreath-Laying Ceremony, The Arlington County Public Safety Combined Honor Guard will perform a Presentation of Colors and lay a wreath at county government headquarters at 2100 Clarendon Blvd at 9 a.m tomorrow (Friday). A moment of silence will be observed at 9:37 a.m., when American Airlines Flight 77 struck the Pentagon., The ceremony can be steamed on the county website, YouTube, or Facebook, or viewed on Comcast channel 1085 or Verizon FiOS channel 39., Memorial 5K, Arlington’s police and fire departments, the Sheriff’s Office and the Emergency Communications Center will host the annual 9/11 Memorial 5K Run and fundraiser this Saturday. The race starts at 6 p.m. on Saturday, Sept. 11 at the DoubleTree by Hilton Hotel in Crystal City (300 Army Navy Drive). The in-person race is at capacity, but the event is still registering virtual participants., Pentagon Memorial Event, The National 9/11 Pentagon Memorial is currently closed due to COVID-19, with no reopening date set. Family members of victims and other invited guests will be admitted this weekend for a seated event with social distancing and various speakers., ‘Ride of Hope’ Cycling Event, Cyclists will ride 15 miles, starting at 7 a.m. on Saturday, and stop at the nine Arlington fire stations that responded to the attack. The ride ends with a moment of silence and a wreath ceremony., Those who responded 20 years ago will ride to the Pentagon 9/11 memorial and lay another wreath. All retired and active first responders are invited, as well as family and friends 18 and older. If spots are available, other adults can join as well., Photography Exhibit, Three local photographers will host a photography exhibit entitled “Still Standing — Still Free” at Fashion Centre at Pentagon City mall, with original photos, a video of the immediate aftermath, 9/11 artifacts and never-before-seen snapshots. The display will be free to the public. It runs from Saturday, Sept. 11 through Monday, Oct. 11., First Responders Cup Tournament, A softball tournament at the Barcroft Park (4200 S. Four Mile Run Drive) on Saturday will raise money for Pentagon Disaster Relief charities. All games are free to participate in and open to anyone who is interested. Opening ceremonies start at 7:30 a.m. Saturday, and will feature the Armed Services Color Guard, the 3rd Army Old Guard Ceremonial Fife and Drums Corps, and the fire department.,
Topic 3:
, This regularly scheduled sponsored Q&A column is written by Eli Tucker, Arlington-based Realtor and Arlington resident. Please submit your questions to him via email for response in future columns. Video summaries of some articles can be found on YouTube on the Ask Eli, Live With Jean playlist. Enjoy!, Question: Is the single-family home market still as intense as it was earlier this year?, Answer: In January I’ll do a deep dive into the 2021 market performance with a focus on home values, but this week I wanted to dig into some key supply and demand metrics for single-family, townhouse and duplex homes in 2021 to highlight how the intensity of the market has shifted over the course of the year., I’m focusing on the single-family, townhouse and duplex (non-condo/apartment-style) market here because that was the market that exploded locally and nationwide in the wake of COVID. It’s important to note, however, when looking at the Arlington market that we didn’t experience nearly the extreme change as many other regional or national markets because things were already competitive thanks in part to Amazon HQ2 and because COVID-based demand tended to favor less expensive markets and markets that offered more space (land and house)., The trends for Arlington can be summarized below, highlighted by charts to follow:, Supply, Demand, The charts below highlight my supply and demand findings. A few notes on the data that makes up the charts:, The Market Moved Quickly, Gave Buyers Little Time to Think, Many buyers were forced to make significant purchase decisions in a matter of hours or even sight unseen to secure a good home. During peak spring demand, less than 20% of homes listed for sale sat on the market for more than two weeks and nearly 60% went under contract in less than one week., , Most Buyers Paid Over Asking Price, On average, buyers paid .2% over the asking price this year and for those who went under contract during a home’s first week on the market, the average buyer paid 2.8% over asking, peaking at an average of 5% over ask in the 9th Period (homes listed April 18-May 1). Remember, these are averages, there were plenty of people paying significantly more than that over the asking price., Things have gotten slightly more manageable for buyers in the 2nd half of the year with a lot more homes selling at or below asking price, but even with tapering demand, buyers in the 2nd half of the year who go under contract in the first two weeks a home was listed paid an average of 1.5% over ask., , Supply Unusually High in 2nd Half, Average Days On Market Increasing, As noted above in my summary, supply volume broke familiar seasonal trends with a consistently strong flow of listings coming to market through the 2nd half and even into Q4. Thus, slightly less demand and unusually high new supply has led to modest increases in average days on market and less fierce competition., , If you’d like to discuss buying, selling, investing, or renting, don’t hesitate to reach out to me at [email protected]., If you’d like a question answered in my weekly column or to discuss buying, selling, renting, or investing, please send an email to [email protected]. To read any of my older posts, visit the blog section of my website at EliResidential.com. Call me directly at 703-539-2529., Video summaries of some articles can be found on YouTube on the Ask Eli, Live With Jean playlist., Eli Tucker is a licensed Realtor in Virginia, Washington DC, and Maryland with RLAH Real Estate, 4040 N Fairfax Dr #10C Arlington VA 22203. 703-390-9460.
Topic 4:
, This sponsored column is by James Montana, Esq. and Doran Shemin, Esq., practicing attorneys at Steelyard LLC, an immigration-focused law firm located in Arlington, Virginia. The legal information given here is general in nature. If you want legal advice, contact James for an appointment., , COVID-19 has changed the way we do business. We’ve ended all client visits at our office, and when green cards arrive, we deliver them in person. We are also doing our part to help out in the community, by volunteering with the Medical Reserve Corps (check out James’ new gear!) and by helping other local businesses apply for Paycheck Protection Program funding., Immigration is our specialty, though, so we want to provide the latest information here about what parts of the immigration system are operational and which are not. As always, consult your lawyer if you have questions about your own particular circumstances., U.S. Citizenship and Immigration Services (USCIS) is still processing green cards, work permits, asylum applications and other paper-based requests. But the doors are shut to in-person visits. As of March 18, all in-person services, including naturalization interviews, citizenship oath ceremonies and asylum interviews are canceled through May 3., Application Support Centers, which process applicants’ fingerprints for various benefit applications, are also closed through May 3. However, the good news for work permit applicants is that if the applicant has previously provided fingerprints, USCIS will use the previously submitted fingerprints to continue to process the application., USCIS has recognized that during the COVID-19 crisis, it may be more difficult to obtain certain documentation to respond to a request for more evidence or file an appeal. Therefore, USCIS will accept responses for up to 60 days after the original due date for any response or appeal issued or due between March 1 and May 1., USCIS also acknowledged that it is safer for clients and attorneys to avoid meeting in person. Normally, USCIS requires wet ink signatures on many applications and petitions. In light of the COVID-19 crisis, USCIS is temporarily accepting scanned or photocopied signatures so attorneys and applicants do not have to hand paperwork back and forth or meet to sign documents., Customs and Border Protection, along with USCIS, is also assisting foreign travelers. Many people can come to the United States for a period of 90 days without a visa based on the Visa Waiver Program. However, COVID-19 has left many travelers stranded and unable to leave the United States before the 90 days runs out. Customs and Border Protection and USCIS are assisting travelers obtain a “satisfactory departure” and 30-day extension in the hope that the inability to leave does not negatively impact future travel., The immigration courts have also taken steps to protect the public. All court hearings for immigrants who are not detained have been suspended through May 1 and the courts will reschedule the hearing for a later date. This step is important because on any given morning, about 150 people can pass through a single courtroom for a non-detained hearing., Further, many courts have standing orders that allow attorneys to appear telephonically without requesting to do so in advance and without filing a motion for hearings involving a detained immigrant. All of the immigration judges at the Baltimore and Arlington Immigration Courts are allowing attorneys to appear telephonically., Some courts have also allowed attorneys to file documents via email instead of filing in person or by mail. The Executive Office for Immigration Review has also started sending automated email updates to attorneys registered with the online immigration court portal regarding court closures or delayed openings., Finally, the Department of State has suspended all routine visa services unless there is a true emergency. However, the Department of State continues to process some H-2 visas, which are largely issued to seasonal agricultural workers, due to the importance of maintaining the food supply chain during this crisis. U.S. citizen services also continue to be available., Our office recognizes that we are all facing unique challenges in light of the COVID-19 pandemic, and we appreciate the various agencies’ efforts to flatten the curve. We are also doing our part to keep the community safe while continuing to assist our clients by practicing social distancing measures and adhering to the Virginia stay at home order., If you or someone you know is concerned about the effect of COVID-19 on their immigration case, we are still here to help. As always, we also welcome any comments and will do our best to respond.
Topic 5:
Locals looking for a good fish fry this Lenten season will have to cast their nets outside Arlington., People won’t have to go far to indulge for Fat Tuesday — which is today — whether that’s with King Cake from Bayou Bakery or Cajun food at Ragtime. But getting to a fish fry may involve a drive into Falls Church or Fairfax County., Catholics and some other Christian sects fast on Ash Wednesday (tomorrow) and certain days during Lent, the 40-day period leading up to Easter. Traditionally, that involves abstaining from flesh meats, such as chicken, beef or pork, on Fridays., Over time, the fasting tradition turned into the church fish fry, often run by a local Knights of Columbus chapter to benefit charity or a parish to support their various ministries. The menu typically includes baked and fried fish, French fries, coleslaw, mac and cheese, other assorted sides and dessert., But Midwest and Northeastern transplants to the D.C. area have noted their beloved fish fries aren’t as popular in and around D.C., “I found that fish fries are mostly up north, as I have a lot of family up that way,” says Myles McMorrow, who sits on the board of Arlington’s chapter of the Knights of Columbus on Little Falls Road. “[For] me, personally, I have never heard of a fish fry in the D.C. metro area and I grew up here.”, He says the local Knights observe Lent by dropping meaty meals from its council restaurant’s menu. Some local churches in the Diocese of Arlington host meatless soup suppers, including St. Agnes Catholic Church in Arlington., Those who are Catholic, curious or culturally homesick are told their best bet for finding a fish fry is to drive deeper into Virginia., Fish fries are mostly a Midwestern and Rust Belt phenomenon because European Catholic immigrants relied on the abundant fish of the Great Lakes to observe their religious fasts. Over time, the tradition may have blended with an African-American tradition of gathering together for fish fries, which began on plantations and continued after Emancipation as families moved North., Churches in the Diocese of Arlington had to sacrifice Lenten gatherings in 2020. In 2021, options were sparse, but this year, a number of parishes have resurrected fish fries and soup suppers., The closest for Arlingtonians is hosted by St. James Catholic Church in Falls Church. It was started in 2010 by a group of parishioners that included a homesick Ohio native., Every year, hand-battered fish and scratch-made potatoes, hush puppies, coleslaw and carrot cake reel in pilgrims from D.C. to Fredericksburg. People can buy T-shirts emblazoned with the year’s slogan, which is always a fish pun. (This year’s is that the 13th annual fish fry “is trout of this world.”), “I remember this couple who drove in from D.C.,” says parishioner Karen Bushaw-Newton. “They said, ‘We just heard there was a fish fry and we came to check it out.’ We know a lot of the parishioners who come — and we have a lot of regulars and families — [and] we have people like that couple who just wanted to see what a fish fry was like.”, When COVID-19 hit, the fish fry turned into a drive-thru that, on some Fridays, served more than 1,000 meals in three hours., “I highly encourage anyone and everybody to come. We don’t ask your faith when you come in the door — it’s just a way to come celebrate,” Bushaw-Newton said., For those who want to go farther afield, there are a number of other Northern Virginia fish fries, though each would require a longer drive in Friday rush hour traffic. Below are a few of the options., Read More
Topic 6:
It’s less than a week before Christmas and Moore’s Barbershop is bustling., Mask-wearing barbers are clipping, trimming, and shaving hair, while several customers wait for their chance in the chair at the small shop on Langston Blvd. There’s an echo of chatter, conversations ranging from politics to football to a mutual friend who got a new job., By the window stands Jim Moore Jr., the owner, cutting and chatting at the same time. It was in 1960, when his father — Jim Moore Sr. — opened this shop in the Halls Hill neighborhood to cater to Arlington’s Black community, who were often not welcome in white barbers’ chairs., For more than six decades, the shop has thrived as a focal point for the community, a place where all were welcome and lifelong friendships have formed., But on Nov. 7, its patriarch Jim Moore Sr. died at the age of 88., Today, James Thomas Moore Sr. Transitioned into his greater self. Mr. Moore started Mr. Moore’s barber shop in 1960 and “started” me three years later. His example helped me and countless other become better people. I love you dad and will always miss you ❤️!#dmv #barber pic.twitter.com/ppTCyohOGJ, — James Moore (@Mooresbarber) November 8, 2021, , Now, several weeks since his death, memories are fluttering down much like hair trimmings from a fresh cut., “Always jovial,” says Keaton Hopkins describing the elder Moore. Hopkins has been getting his haircut here for more than thirty years, since he was five years old. “Always smiling… We always had a great conversation.”, “He never seemed to have a bad day,” says Clay Pinson, a barber at the shop for about twenty years. “He was always in a good mood.”, His son, Jim, notes that these are common refrains, that his father was kind, a good conversationalist, and knew how to make people feel special., “People have kept coming to me since his passing to tell me stories of the things he’s done for them and the lessons they learned from him,” Moore Jr. tells ARLnow, emotion coming through his voice. “That’s just who he was. He made a difference for a lot of people.”, Moore Sr. was born in North Carolina, served in the Korean War, and went to barber school before finding his way to Arlington, after getting a tip that the Halls Hill neighborhood was in need of a barber’s services. While there were Black barbers in the county and nearby in D.C., white clients would only go to them if the clippers and scissors had not been used on a Black client., “They refused to cut Black people’s hair,” says Moore Jr., So, Moore Sr. opened his own shop with a partner, Rudolf Becton, and ingrained himself in the community. In addition to being a barber, he was also a volunteer firefighter at the nearby, historic Fire Station #8. In 1962, Jim Moore Jr., was born and it didn’t take long before the young son went to work at the family business., “I started when I was seven [years old] and my job was cleaning it up for him, sweeping hair,” he says. “I didn’t start cutting hair until I was a teenager.”, He also followed in his father’s footsteps by becoming an Arlington firefighter, serving the county for more than thirty years before retiring in 2020. On his off-days from the department, though, he would stand by his father’s side., Moore Jr. learned that being a barber is about so much more than just knowing how to handle scissors. The profession requires listening, building relationships, and making people feel comfortable., “Cutting hair is an intimate activity,” says the younger Moore. “You are close to somebody, you touch them, you smell them. You can see the sweat and tension when they are talking about certain subjects. You need to know how to read a person.”, And there was no one better at those skills than the elder Moore., “I called it his superpower. The ability to… allow people the space to be their authentic self,” Moore Jr. says., Throughout its history, Moore’s Barbershop has continued to be a place for everyone. In fact, it’s often cited as the first integrated barber shop in Arlington. Moore Jr. says his father never believed in segregation, knowing that a good haircut and great conversation were universal desires., Moore Jr. has continued this tradition of providing for the community, including giving away books to kids, free back-to-school haircuts, and simply by taking the load off of beleaguered spouses., “What my dad taught me is that you can be successful in many ways. It doesn’t have to be a great big billion dollar house or a great big million dollar company,” says Moore Jr. “The smallest things can make a huge difference. That’s what he always put out there.”, Read More
Topic 7:
, (updated at 3:35 p.m.) The Arlington School Board is suing to stop Gov. Glenn Youngkin’s executive order that doesn’t allow school systems to require students to wear masks., The lawsuit filed this morning (Monday) challenges the order issued by Youngkin on Jan. 15, his first day in office. Arlington joined school boards from Fairfax County, Alexandria City, Falls Church City, Hampton City, Prince William County and the City of Richmond in the suit., The order states parents should be able to “elect for their children not to be subject to any mask mandate in effect at the child’s school or educational program.”, The order was supposed to take effect today but school districts across the state, including Arlington, already made decisions at the local level to go against the order and keep a mask requirement in place as part of a strategy to reduce the spread of Covid and maintain in-person instruction., The lawsuit challenges the constitutionality of the executive order, and defends the right of school boards to enact policy at the local level. The lawsuit also claims the executive order goes against Senate Bill 1303, which was adopted in the General Assembly’s 2021 special session. The law states school boards should follow the Centers for Disease Control and Prevention’s health and safety requirements., “Everyone in our community plays a role in keeping schools open and safe for students through consistent mask wearing and other mitigation measures,” APS Superintendent Fransisco Durán wrote in an email to families. “Our shared goal remains to make sure every student continues to access in-person learning five days per week. We look forward to the opportunity to ease these requirements in APS once public health guidance indicates it is safe to do so.”, APS spokesman Frank Bellavia said the schools continue to follow the same guidelines in place since the beginning of the school year., “If a student is not wearing a mask, our schools are advised to speak to the student and provide them a mask to wear,” he said., He said the vast majority of APS families support and adhere to the health and safety guidelines and when students arrived at school Monday, there were “very few incidents.”, The Arlington School Board put out a statement as well, stating it “stands together with participating school boards across the Commonwealth to defend our constitutional right to set policies and supervise our local schools. We continue to make decisions that allow us to keep schools open and safe for in-person learning, in accordance with Virginia law SB 1303 and the CDC’s guidance regarding the use of universal masks and other layered prevention strategies.”, Over the last seven days, 467 students and 98 staff members were positive for Covid, according to the school system’s COVID-19 dashboard., The full press release from Arlington Public Schools is below., Today, the Schools Boards of Alexandria City, Arlington County, City of Richmond, Fairfax County, Falls Church City, Hampton City and Prince William County, filed a lawsuit to challenge the constitutionality of Executive Order 2 issued by the governor on January 15, 2022. The legal action, representing over 350,000 students across the state, defends the right of school boards to enact policy at the local level, including policies that protect the health and well-being of all students and staff., This legal action centers on fundamental questions about the framework of public education in Virginia, as set out in the Virginia Constitution and by the General Assembly. At issue is whether locally elected school boards have the exclusive authority and responsibility conferred upon them by Article VIII, § 7 of the Constitution of Virginia over supervision of the public schools in their respective communities, or whether an executive order can unilaterally override that constitutional authority., Also at issue is whether a governor can, through executive order, without legislative action by the Virginia General Assembly, reverse a lawfully-adopted statute. In this case, Senate Bill 1303, adopted with the goal of returning students to safe in-person instruction five days a week in March 2021 and still legally in effect, provides that local school boards should follow The Centers for Disease Control and Prevention (CDC) health and safety requirements., Without today’s action, school boards are placed in a legally untenable position — faced with an executive order that is in conflict with the constitution and state law. Today’s action is not politically motivated. These seven school divisions would welcome the opportunity to collaborate with the governor to ensure the safety and welfare of all students., This lawsuit is not brought out of choice, but out of necessity., With COVID-19 transmission rates high, our hospitals at crisis level, and the continued recommendation of health experts to retain universal mask-wearing for the time being, this is simply not the time to remove this critical component of layered health and safety mitigation strategies. School divisions need to continue to preserve their authority to protect and serve all our students, including our most vulnerable, who need these mitigation measures perhaps more than anyone to be able to continue to access in-person instruction.
Code
# structural topic model# I will leverage the time_tag variable included in the dataset as a covariate in my estimates of topical prevalence, I expect some topics to be more prevalent by different dates (expect topics to change over time)# specify model I chose k=7 using the searchK function, shown below/later in this documentmyModel <-stm(arlnow_covid_dfm_new,K =7,prevalence =~ time_year,data = arlnow_covid,max.em.its =1000,seed =1234,init.type ="Spectral")
# plotting out the top topics (as groups of words associated with that topic) and their estimated frequency across the corpusplot(myModel, type ="summary")
Code
# these topics seem to make sense compared to eachother and lda and what I might expect# specify model again but without prevalence = ~ time_yearmyModel2 <-stm(arlnow_covid_dfm_new,K =7,data = arlnow_covid,max.em.its =1000,seed =1234,init.type ="Spectral")
# plotting out the top topics (as groups of words associated with that topic) and their estimated frequency across the corpusplot(myModel2, type ="summary")
Code
# these topics also seem to make sense, generally# choosing k - searchK() lets you estimate a series of different models, then you can plot a series of different evaluation metrics across those choicesdifferentKs <-searchK(arlnow_covid_dfm_new,K =c(3, 5, 7, 10, 20, 30),prevalence =~ time_year,N =200,data = arlnow_covid,max.em.its =1000,init.type ="Spectral")
Beginning Spectral Initialization
Calculating the gram matrix...
Using only 10000 most frequent terms during initialization...
Finding anchor words...
...
Recovering initialization...
....................................................................................................
Initialization complete.
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 1 (approx. per word bound = -8.259)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 2 (approx. per word bound = -7.950, relative change = 3.732e-02)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 3 (approx. per word bound = -7.899, relative change = 6.476e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 4 (approx. per word bound = -7.884, relative change = 1.821e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 5 (approx. per word bound = -7.877, relative change = 1.007e-03)
Topic 1: cases, covid, new, week, said
Topic 2: said, community, board, year, people
Topic 3: school, covid-19, health, new, said
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 6 (approx. per word bound = -7.871, relative change = 6.710e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 7 (approx. per word bound = -7.867, relative change = 5.205e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 8 (approx. per word bound = -7.864, relative change = 3.554e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 9 (approx. per word bound = -7.862, relative change = 2.583e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 10 (approx. per word bound = -7.861, relative change = 2.117e-04)
Topic 1: cases, covid, new, week, said
Topic 2: said, community, business, board, year
Topic 3: school, covid-19, aps, students, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 11 (approx. per word bound = -7.859, relative change = 1.904e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 12 (approx. per word bound = -7.858, relative change = 1.527e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 13 (approx. per word bound = -7.857, relative change = 1.331e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 14 (approx. per word bound = -7.856, relative change = 1.053e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 15 (approx. per word bound = -7.855, relative change = 8.778e-05)
Topic 1: cases, covid, new, week, said
Topic 2: said, community, business, board, people
Topic 3: school, aps, covid-19, students, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 16 (approx. per word bound = -7.855, relative change = 9.116e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 17 (approx. per word bound = -7.854, relative change = 9.327e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 18 (approx. per word bound = -7.853, relative change = 6.621e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 19 (approx. per word bound = -7.853, relative change = 6.486e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 20 (approx. per word bound = -7.852, relative change = 8.673e-05)
Topic 1: cases, covid, new, week, said
Topic 2: said, community, business, people, new
Topic 3: school, aps, students, covid-19, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 21 (approx. per word bound = -7.852, relative change = 7.142e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 22 (approx. per word bound = -7.851, relative change = 4.127e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 23 (approx. per word bound = -7.851, relative change = 3.550e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 24 (approx. per word bound = -7.851, relative change = 4.593e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 25 (approx. per word bound = -7.850, relative change = 5.436e-05)
Topic 1: cases, covid, new, week, said
Topic 2: said, community, business, people, new
Topic 3: school, aps, students, covid-19, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 26 (approx. per word bound = -7.850, relative change = 5.027e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 27 (approx. per word bound = -7.850, relative change = 5.103e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 28 (approx. per word bound = -7.849, relative change = 5.116e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 29 (approx. per word bound = -7.849, relative change = 4.121e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 30 (approx. per word bound = -7.849, relative change = 3.298e-05)
Topic 1: cases, covid, new, week, said
Topic 2: said, community, business, people, new
Topic 3: school, aps, students, covid-19, said
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 31 (approx. per word bound = -7.848, relative change = 2.870e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 32 (approx. per word bound = -7.848, relative change = 2.561e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 33 (approx. per word bound = -7.848, relative change = 1.787e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 34 (approx. per word bound = -7.848, relative change = 1.923e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 35 (approx. per word bound = -7.848, relative change = 2.861e-05)
Topic 1: cases, covid, new, week, said
Topic 2: said, community, business, people, new
Topic 3: school, aps, students, covid-19, said
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 36 (approx. per word bound = -7.847, relative change = 2.611e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 37 (approx. per word bound = -7.847, relative change = 3.008e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 38 (approx. per word bound = -7.847, relative change = 3.091e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 39 (approx. per word bound = -7.847, relative change = 2.506e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 40 (approx. per word bound = -7.846, relative change = 3.208e-05)
Topic 1: cases, covid, new, week, said
Topic 2: said, community, business, new, people
Topic 3: school, aps, students, covid-19, said
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 41 (approx. per word bound = -7.846, relative change = 3.446e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 42 (approx. per word bound = -7.846, relative change = 1.453e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 43 (approx. per word bound = -7.846, relative change = 1.069e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 44 (approx. per word bound = -7.846, relative change = 1.433e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 45 (approx. per word bound = -7.846, relative change = 1.955e-05)
Topic 1: cases, covid, new, week, said
Topic 2: said, community, business, new, people
Topic 3: school, aps, students, covid-19, said
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 46 (approx. per word bound = -7.846, relative change = 2.105e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 47 (approx. per word bound = -7.845, relative change = 2.299e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 48 (approx. per word bound = -7.845, relative change = 1.721e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Model Converged
Beginning Spectral Initialization
Calculating the gram matrix...
Using only 10000 most frequent terms during initialization...
Finding anchor words...
.....
Recovering initialization...
....................................................................................................
Initialization complete.
..................................................................................................................
Completed E-Step (1 seconds).
Completed M-Step.
Completing Iteration 1 (approx. per word bound = -8.251)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 2 (approx. per word bound = -7.795, relative change = 5.522e-02)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 3 (approx. per word bound = -7.736, relative change = 7.513e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 4 (approx. per word bound = -7.722, relative change = 1.876e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 5 (approx. per word bound = -7.713, relative change = 1.107e-03)
Topic 1: cases, covid, new, week, health
Topic 2: said, year, community, board, new
Topic 3: school, covid-19, aps, students, vaccine
Topic 4: said, local, people, time, business
Topic 5: said, says, business, new, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 6 (approx. per word bound = -7.707, relative change = 7.746e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 7 (approx. per word bound = -7.703, relative change = 5.840e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 8 (approx. per word bound = -7.700, relative change = 4.136e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 9 (approx. per word bound = -7.697, relative change = 3.353e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 10 (approx. per word bound = -7.695, relative change = 2.784e-04)
Topic 1: cases, covid, new, week, health
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, covid-19, schools
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 11 (approx. per word bound = -7.693, relative change = 2.660e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 12 (approx. per word bound = -7.691, relative change = 2.413e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 13 (approx. per word bound = -7.690, relative change = 1.922e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 14 (approx. per word bound = -7.688, relative change = 1.700e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 15 (approx. per word bound = -7.687, relative change = 1.344e-04)
Topic 1: cases, covid, new, week, health
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 16 (approx. per word bound = -7.686, relative change = 1.345e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 17 (approx. per word bound = -7.685, relative change = 1.529e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 18 (approx. per word bound = -7.684, relative change = 1.602e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 19 (approx. per word bound = -7.683, relative change = 1.370e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 20 (approx. per word bound = -7.682, relative change = 1.250e-04)
Topic 1: cases, covid, new, week, health
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, business, new, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 21 (approx. per word bound = -7.681, relative change = 1.289e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 22 (approx. per word bound = -7.680, relative change = 1.035e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 23 (approx. per word bound = -7.679, relative change = 7.023e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 24 (approx. per word bound = -7.679, relative change = 8.766e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 25 (approx. per word bound = -7.678, relative change = 7.987e-05)
Topic 1: cases, covid, new, health, week
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 26 (approx. per word bound = -7.678, relative change = 5.397e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 27 (approx. per word bound = -7.677, relative change = 4.322e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 28 (approx. per word bound = -7.677, relative change = 3.813e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 29 (approx. per word bound = -7.677, relative change = 4.755e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 30 (approx. per word bound = -7.676, relative change = 5.681e-05)
Topic 1: cases, covid, new, health, week
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, business, new, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 31 (approx. per word bound = -7.676, relative change = 5.098e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 32 (approx. per word bound = -7.676, relative change = 3.957e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 33 (approx. per word bound = -7.675, relative change = 4.225e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 34 (approx. per word bound = -7.675, relative change = 5.732e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 35 (approx. per word bound = -7.674, relative change = 5.423e-05)
Topic 1: cases, covid, new, health, week
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 36 (approx. per word bound = -7.674, relative change = 3.436e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 37 (approx. per word bound = -7.674, relative change = 2.807e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 38 (approx. per word bound = -7.674, relative change = 3.489e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 39 (approx. per word bound = -7.673, relative change = 4.203e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 40 (approx. per word bound = -7.673, relative change = 2.859e-05)
Topic 1: cases, covid, new, health, week
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 41 (approx. per word bound = -7.673, relative change = 1.936e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 42 (approx. per word bound = -7.673, relative change = 1.752e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 43 (approx. per word bound = -7.673, relative change = 1.702e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 44 (approx. per word bound = -7.673, relative change = 2.324e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 45 (approx. per word bound = -7.672, relative change = 3.245e-05)
Topic 1: cases, covid, new, health, week
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 46 (approx. per word bound = -7.672, relative change = 2.725e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 47 (approx. per word bound = -7.672, relative change = 2.664e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 48 (approx. per word bound = -7.672, relative change = 2.773e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 49 (approx. per word bound = -7.671, relative change = 3.654e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 50 (approx. per word bound = -7.671, relative change = 4.596e-05)
Topic 1: cases, covid, new, health, week
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 51 (approx. per word bound = -7.671, relative change = 2.886e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 52 (approx. per word bound = -7.671, relative change = 1.995e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 53 (approx. per word bound = -7.671, relative change = 2.721e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 54 (approx. per word bound = -7.670, relative change = 4.847e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 55 (approx. per word bound = -7.670, relative change = 4.415e-05)
Topic 1: cases, covid, new, health, week
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 56 (approx. per word bound = -7.670, relative change = 2.074e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 57 (approx. per word bound = -7.669, relative change = 1.851e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 58 (approx. per word bound = -7.669, relative change = 2.210e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 59 (approx. per word bound = -7.669, relative change = 2.924e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 60 (approx. per word bound = -7.669, relative change = 3.561e-05)
Topic 1: cases, covid, new, health, week
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 61 (approx. per word bound = -7.669, relative change = 2.599e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 62 (approx. per word bound = -7.668, relative change = 2.221e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 63 (approx. per word bound = -7.668, relative change = 2.000e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 64 (approx. per word bound = -7.668, relative change = 1.568e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 65 (approx. per word bound = -7.668, relative change = 1.402e-05)
Topic 1: cases, covid, new, health, week
Topic 2: said, community, year, board, million
Topic 3: school, aps, students, schools, covid-19
Topic 4: said, local, people, time, business
Topic 5: said, says, new, business, restaurant
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 66 (approx. per word bound = -7.668, relative change = 2.394e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 67 (approx. per word bound = -7.668, relative change = 2.953e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 68 (approx. per word bound = -7.668, relative change = 1.459e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Model Converged
Beginning Spectral Initialization
Calculating the gram matrix...
Using only 10000 most frequent terms during initialization...
Finding anchor words...
.......
Recovering initialization...
....................................................................................................
Initialization complete.
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 1 (approx. per word bound = -8.237)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 2 (approx. per word bound = -7.677, relative change = 6.791e-02)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 3 (approx. per word bound = -7.616, relative change = 8.018e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 4 (approx. per word bound = -7.603, relative change = 1.663e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 5 (approx. per word bound = -7.597, relative change = 8.356e-04)
Topic 1: cases, covid, new, week, health
Topic 2: said, board, new, year, community
Topic 3: school, covid, vaccine, covid-19, aps
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, city, new, school
Topic 6: said, people, can, public, board
Topic 7: covid-19, business, arts, said, pandemic
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 6 (approx. per word bound = -7.592, relative change = 6.248e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 7 (approx. per word bound = -7.588, relative change = 4.674e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 8 (approx. per word bound = -7.586, relative change = 3.701e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 9 (approx. per word bound = -7.584, relative change = 2.650e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 10 (approx. per word bound = -7.582, relative change = 1.932e-04)
Topic 1: cases, covid, new, week, rate
Topic 2: said, board, year, new, community
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, city, new, pentagon
Topic 6: said, people, can, public, board
Topic 7: covid-19, business, arts, said, pandemic
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 11 (approx. per word bound = -7.581, relative change = 1.672e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 12 (approx. per word bound = -7.580, relative change = 1.600e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 13 (approx. per word bound = -7.578, relative change = 1.579e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 14 (approx. per word bound = -7.577, relative change = 1.325e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 15 (approx. per word bound = -7.577, relative change = 1.157e-04)
Topic 1: cases, covid, new, week, rate
Topic 2: said, year, new, board, park
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, city, new, pentagon
Topic 6: said, people, can, public, board
Topic 7: covid-19, business, arts, said, pandemic
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 16 (approx. per word bound = -7.576, relative change = 9.822e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 17 (approx. per word bound = -7.575, relative change = 8.548e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 18 (approx. per word bound = -7.575, relative change = 6.454e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 19 (approx. per word bound = -7.574, relative change = 5.010e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 20 (approx. per word bound = -7.574, relative change = 4.582e-05)
Topic 1: cases, covid, new, week, rate
Topic 2: said, new, year, board, park
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, city, new, street
Topic 6: said, people, can, board, public
Topic 7: covid-19, business, arts, said, pandemic
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 21 (approx. per word bound = -7.574, relative change = 4.925e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 22 (approx. per word bound = -7.573, relative change = 6.182e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 23 (approx. per word bound = -7.573, relative change = 6.563e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 24 (approx. per word bound = -7.572, relative change = 5.386e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 25 (approx. per word bound = -7.572, relative change = 4.037e-05)
Topic 1: cases, covid, new, week, rate
Topic 2: said, new, board, year, park
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, new, city, street
Topic 6: said, people, can, board, public
Topic 7: covid-19, business, arts, pandemic, said
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 26 (approx. per word bound = -7.572, relative change = 4.086e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 27 (approx. per word bound = -7.571, relative change = 5.520e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 28 (approx. per word bound = -7.571, relative change = 5.365e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 29 (approx. per word bound = -7.570, relative change = 4.551e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 30 (approx. per word bound = -7.570, relative change = 4.507e-05)
Topic 1: cases, covid, new, week, rate
Topic 2: said, new, park, community, board
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, new, city, street
Topic 6: said, people, can, board, public
Topic 7: covid-19, business, arts, pandemic, said
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 31 (approx. per word bound = -7.570, relative change = 3.887e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 32 (approx. per word bound = -7.570, relative change = 3.027e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 33 (approx. per word bound = -7.569, relative change = 3.214e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 34 (approx. per word bound = -7.569, relative change = 4.158e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 35 (approx. per word bound = -7.569, relative change = 4.887e-05)
Topic 1: cases, covid, new, week, rate
Topic 2: said, new, park, community, board
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, new, city, street
Topic 6: said, people, board, can, public
Topic 7: covid-19, business, arts, pandemic, million
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 36 (approx. per word bound = -7.568, relative change = 6.124e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 37 (approx. per word bound = -7.568, relative change = 7.774e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 38 (approx. per word bound = -7.567, relative change = 9.833e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 39 (approx. per word bound = -7.566, relative change = 9.139e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 40 (approx. per word bound = -7.566, relative change = 4.888e-05)
Topic 1: cases, covid, new, week, rate
Topic 2: said, new, park, community, year
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, new, street, city
Topic 6: said, people, board, can, public
Topic 7: covid-19, business, arts, pandemic, million
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 41 (approx. per word bound = -7.566, relative change = 3.718e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 42 (approx. per word bound = -7.565, relative change = 3.882e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 43 (approx. per word bound = -7.565, relative change = 4.689e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 44 (approx. per word bound = -7.565, relative change = 4.687e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 45 (approx. per word bound = -7.564, relative change = 5.457e-05)
Topic 1: cases, covid, new, week, rate
Topic 2: said, new, park, community, year
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, new, street, city
Topic 6: said, people, board, can, public
Topic 7: covid-19, business, arts, pandemic, million
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 46 (approx. per word bound = -7.564, relative change = 4.750e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 47 (approx. per word bound = -7.564, relative change = 2.406e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 48 (approx. per word bound = -7.563, relative change = 2.180e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 49 (approx. per word bound = -7.563, relative change = 2.505e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 50 (approx. per word bound = -7.563, relative change = 1.881e-05)
Topic 1: cases, covid, new, week, rate
Topic 2: said, new, community, park, year
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, new, street, city
Topic 6: said, people, board, can, public
Topic 7: covid-19, business, arts, pandemic, million
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 51 (approx. per word bound = -7.563, relative change = 1.358e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 52 (approx. per word bound = -7.563, relative change = 1.693e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 53 (approx. per word bound = -7.563, relative change = 2.151e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 54 (approx. per word bound = -7.563, relative change = 2.237e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 55 (approx. per word bound = -7.562, relative change = 1.543e-05)
Topic 1: cases, covid, new, week, rate
Topic 2: said, new, community, park, year
Topic 3: school, aps, students, public, schools
Topic 4: local, said, time, arlnow, people
Topic 5: said, says, new, street, city
Topic 6: said, people, board, can, public
Topic 7: covid-19, business, arts, pandemic, million
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Model Converged
Beginning Spectral Initialization
Calculating the gram matrix...
Using only 10000 most frequent terms during initialization...
Finding anchor words...
..........
Recovering initialization...
....................................................................................................
Initialization complete.
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 1 (approx. per word bound = -8.196)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 2 (approx. per word bound = -7.521, relative change = 8.234e-02)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 3 (approx. per word bound = -7.461, relative change = 7.990e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 4 (approx. per word bound = -7.447, relative change = 1.826e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 5 (approx. per word bound = -7.440, relative change = 8.809e-04)
Topic 1: week, said, local, weekend, last
Topic 2: said, board, new, people, community
Topic 3: school, aps, students, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, school, city, open
Topic 6: said, people, health, public, work
Topic 7: covid-19, business, arts, million, said
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, said, event, park, new
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 6 (approx. per word bound = -7.436, relative change = 5.833e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 7 (approx. per word bound = -7.433, relative change = 3.590e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 8 (approx. per word bound = -7.431, relative change = 2.838e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 9 (approx. per word bound = -7.429, relative change = 2.592e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 10 (approx. per word bound = -7.428, relative change = 1.888e-04)
Topic 1: week, said, local, weekend, last
Topic 2: said, board, new, people, community
Topic 3: school, aps, students, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, city, restaurant, open
Topic 6: said, people, health, public, work
Topic 7: covid-19, business, arts, million, businesses
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, said, park, community
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 11 (approx. per word bound = -7.427, relative change = 1.322e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 12 (approx. per word bound = -7.426, relative change = 1.129e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 13 (approx. per word bound = -7.426, relative change = 8.521e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 14 (approx. per word bound = -7.425, relative change = 7.147e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 15 (approx. per word bound = -7.425, relative change = 5.527e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, city, project, restaurant
Topic 6: said, people, health, work, public
Topic 7: covid-19, business, arts, businesses, million
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, said, park, community
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 16 (approx. per word bound = -7.424, relative change = 5.372e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 17 (approx. per word bound = -7.424, relative change = 5.058e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 18 (approx. per word bound = -7.423, relative change = 5.414e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 19 (approx. per word bound = -7.423, relative change = 5.759e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 20 (approx. per word bound = -7.423, relative change = 6.213e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, project, city, street
Topic 6: said, people, health, work, public
Topic 7: covid-19, business, arts, businesses, million
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, said, community, park
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 21 (approx. per word bound = -7.422, relative change = 6.448e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 22 (approx. per word bound = -7.422, relative change = 6.715e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 23 (approx. per word bound = -7.421, relative change = 8.606e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 24 (approx. per word bound = -7.420, relative change = 7.212e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 25 (approx. per word bound = -7.420, relative change = 5.792e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, project, street, open
Topic 6: said, people, health, work, public
Topic 7: business, covid-19, arts, businesses, pandemic
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, said, community, park
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 26 (approx. per word bound = -7.420, relative change = 6.056e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 27 (approx. per word bound = -7.419, relative change = 4.526e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 28 (approx. per word bound = -7.419, relative change = 2.851e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 29 (approx. per word bound = -7.419, relative change = 2.635e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 30 (approx. per word bound = -7.419, relative change = 2.779e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, project, street, open
Topic 6: said, people, health, work, public
Topic 7: business, covid-19, arts, businesses, pandemic
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, said, community, park
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 31 (approx. per word bound = -7.418, relative change = 3.144e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 32 (approx. per word bound = -7.418, relative change = 3.312e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 33 (approx. per word bound = -7.418, relative change = 3.895e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 34 (approx. per word bound = -7.417, relative change = 4.566e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 35 (approx. per word bound = -7.417, relative change = 3.576e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, project, street, open
Topic 6: said, people, health, work, public
Topic 7: covid-19, business, arts, businesses, pandemic
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, said, community, park
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 36 (approx. per word bound = -7.417, relative change = 2.291e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 37 (approx. per word bound = -7.417, relative change = 1.736e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 38 (approx. per word bound = -7.417, relative change = 1.504e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 39 (approx. per word bound = -7.417, relative change = 1.592e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 40 (approx. per word bound = -7.417, relative change = 1.902e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, project, street, open
Topic 6: said, people, health, work, public
Topic 7: covid-19, business, arts, businesses, pandemic
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, said, community, park
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 41 (approx. per word bound = -7.416, relative change = 2.350e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 42 (approx. per word bound = -7.416, relative change = 2.789e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 43 (approx. per word bound = -7.416, relative change = 3.026e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 44 (approx. per word bound = -7.416, relative change = 2.360e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 45 (approx. per word bound = -7.416, relative change = 2.227e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, project, street, open
Topic 6: said, people, health, work, public
Topic 7: covid-19, business, arts, businesses, pandemic
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, said, community, park
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 46 (approx. per word bound = -7.415, relative change = 2.492e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 47 (approx. per word bound = -7.415, relative change = 3.595e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 48 (approx. per word bound = -7.415, relative change = 3.973e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 49 (approx. per word bound = -7.415, relative change = 2.882e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 50 (approx. per word bound = -7.414, relative change = 2.460e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, project, street, open
Topic 6: said, people, health, work, public
Topic 7: covid-19, business, arts, businesses, pandemic
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, said, community, park
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 51 (approx. per word bound = -7.414, relative change = 3.152e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 52 (approx. per word bound = -7.414, relative change = 3.897e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 53 (approx. per word bound = -7.414, relative change = 3.200e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 54 (approx. per word bound = -7.414, relative change = 2.702e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 55 (approx. per word bound = -7.413, relative change = 2.989e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, project, street, restaurant
Topic 6: said, people, health, work, public
Topic 7: covid-19, business, arts, businesses, pandemic
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, community, said, new
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 56 (approx. per word bound = -7.413, relative change = 3.895e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 57 (approx. per word bound = -7.413, relative change = 2.105e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 58 (approx. per word bound = -7.413, relative change = 2.413e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 59 (approx. per word bound = -7.412, relative change = 2.652e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 60 (approx. per word bound = -7.412, relative change = 1.758e-05)
Topic 1: week, said, local, weekend, feel
Topic 2: said, board, new, people, community
Topic 3: school, students, aps, schools, said
Topic 4: said, local, arlnow, time, people
Topic 5: said, says, project, street, restaurant
Topic 6: said, people, health, work, public
Topic 7: covid-19, business, arts, businesses, pandemic
Topic 8: market, new, year, said, last
Topic 9: cases, covid, health, new, reported
Topic 10: can, event, community, said, new
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Model Converged
Beginning Spectral Initialization
Calculating the gram matrix...
Using only 10000 most frequent terms during initialization...
Finding anchor words...
....................
Recovering initialization...
....................................................................................................
Initialization complete.
..................................................................................................................
Completed E-Step (1 seconds).
Completed M-Step.
Completing Iteration 1 (approx. per word bound = -8.100)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 2 (approx. per word bound = -7.230, relative change = 1.075e-01)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 3 (approx. per word bound = -7.135, relative change = 1.308e-02)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 4 (approx. per word bound = -7.116, relative change = 2.615e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 5 (approx. per word bound = -7.109, relative change = 1.092e-03)
Topic 1: week, weekend, feel, past, discuss
Topic 2: said, new, community, arlnow, board
Topic 3: test, covid, tests, library, covid-19
Topic 4: local, time, people, arlnow, said
Topic 5: restaurant, says, said, new, open
Topic 6: said, cases, two, people, one
Topic 7: arts, covid-19, community, said, public
Topic 8: day, holiday, new, friday, still
Topic 9: cases, covid, new, reported, rate
Topic 10: said, can, says, just, community
Topic 11: said, board, million, budget, year
Topic 12: vaccine, health, covid-19, cases, covid
Topic 13: business, businesses, small, market, year
Topic 14: can, park, covid-19, also, new
Topic 15: public, health, new, berry, care
Topic 16: people, said, one, transit, use
Topic 17: said, church, community, bus, time
Topic 18: school, aps, students, schools, said
Topic 19: said, group, people, new, office
Topic 20: barstool, freddie's, fund, bar, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 6 (approx. per word bound = -7.105, relative change = 5.372e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 7 (approx. per word bound = -7.102, relative change = 3.285e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 8 (approx. per word bound = -7.101, relative change = 2.169e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 9 (approx. per word bound = -7.100, relative change = 1.084e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 10 (approx. per word bound = -7.100, relative change = 8.084e-05)
Topic 1: week, weekend, said, feel, past
Topic 2: said, new, community, arlnow, board
Topic 3: test, tests, testing, covid, library
Topic 4: local, time, people, arlnow, said
Topic 5: restaurant, says, said, new, open
Topic 6: said, cases, two, people, one
Topic 7: arts, community, said, covid-19, public
Topic 8: day, holiday, market, new, friday
Topic 9: cases, covid, new, reported, rate
Topic 10: said, can, says, just, community
Topic 11: said, board, million, budget, year
Topic 12: vaccine, health, covid-19, cases, covid
Topic 13: business, businesses, small, market, year
Topic 14: can, park, covid-19, also, new
Topic 15: public, health, new, berry, employees
Topic 16: people, said, one, transit, use
Topic 17: said, church, community, bus, time
Topic 18: school, aps, students, schools, said
Topic 19: said, group, office, people, new
Topic 20: barstool, freddie's, fund, bar, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 11 (approx. per word bound = -7.099, relative change = 6.247e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 12 (approx. per word bound = -7.099, relative change = 4.925e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 13 (approx. per word bound = -7.098, relative change = 4.135e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 14 (approx. per word bound = -7.098, relative change = 4.006e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 15 (approx. per word bound = -7.098, relative change = 3.493e-05)
Topic 1: week, weekend, said, feel, anything
Topic 2: said, new, community, arlnow, board
Topic 3: testing, test, tests, covid, library
Topic 4: local, time, people, arlnow, said
Topic 5: restaurant, says, said, new, open
Topic 6: said, cases, two, people, one
Topic 7: arts, community, said, covid-19, public
Topic 8: day, holiday, market, new, friday
Topic 9: cases, covid, new, reported, rate
Topic 10: said, can, says, community, just
Topic 11: said, board, million, budget, year
Topic 12: vaccine, health, covid-19, cases, covid
Topic 13: business, businesses, small, market, year
Topic 14: can, park, covid-19, also, new
Topic 15: public, health, new, berry, care
Topic 16: people, said, one, transit, use
Topic 17: said, church, community, bus, time
Topic 18: school, aps, students, schools, said
Topic 19: said, group, office, people, new
Topic 20: barstool, freddie's, fund, bar, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 16 (approx. per word bound = -7.098, relative change = 2.989e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 17 (approx. per word bound = -7.098, relative change = 2.343e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 18 (approx. per word bound = -7.097, relative change = 2.633e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 19 (approx. per word bound = -7.097, relative change = 2.977e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 20 (approx. per word bound = -7.097, relative change = 2.301e-05)
Topic 1: week, weekend, said, feel, snow
Topic 2: said, new, community, arlnow, board
Topic 3: testing, test, tests, covid, library
Topic 4: local, time, people, arlnow, said
Topic 5: restaurant, says, said, new, open
Topic 6: said, cases, two, people, one
Topic 7: arts, community, said, covid-19, public
Topic 8: day, holiday, market, new, also
Topic 9: cases, covid, new, reported, rate
Topic 10: said, can, says, just, community
Topic 11: said, board, million, budget, year
Topic 12: vaccine, health, covid-19, cases, booster
Topic 13: business, businesses, small, market, grant
Topic 14: can, park, covid-19, also, new
Topic 15: public, health, new, berry, care
Topic 16: people, said, one, transit, use
Topic 17: said, church, community, bus, time
Topic 18: school, aps, students, schools, said
Topic 19: said, group, office, people, new
Topic 20: barstool, freddie's, fund, bar, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 21 (approx. per word bound = -7.097, relative change = 3.095e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 22 (approx. per word bound = -7.096, relative change = 4.595e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 23 (approx. per word bound = -7.096, relative change = 4.915e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 24 (approx. per word bound = -7.096, relative change = 2.979e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 25 (approx. per word bound = -7.096, relative change = 1.363e-05)
Topic 1: week, weekend, snow, said, feel
Topic 2: said, new, community, arlnow, board
Topic 3: testing, test, tests, covid, library
Topic 4: local, time, people, arlnow, said
Topic 5: restaurant, says, said, new, open
Topic 6: said, cases, two, people, one
Topic 7: arts, community, said, covid-19, public
Topic 8: day, holiday, market, new, also
Topic 9: cases, covid, new, reported, rate
Topic 10: said, can, says, just, community
Topic 11: said, board, million, budget, year
Topic 12: vaccine, health, covid-19, cases, booster
Topic 13: business, businesses, small, market, grant
Topic 14: can, park, covid-19, also, new
Topic 15: public, health, new, berry, care
Topic 16: people, said, one, transit, use
Topic 17: said, church, community, bus, time
Topic 18: school, aps, students, schools, said
Topic 19: said, group, office, new, people
Topic 20: barstool, freddie's, fund, bar, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 26 (approx. per word bound = -7.096, relative change = 1.219e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 27 (approx. per word bound = -7.096, relative change = 1.228e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 28 (approx. per word bound = -7.096, relative change = 1.421e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 29 (approx. per word bound = -7.095, relative change = 1.418e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 30 (approx. per word bound = -7.095, relative change = 1.286e-05)
Topic 1: week, snow, weekend, said, feel
Topic 2: said, new, community, arlnow, board
Topic 3: testing, test, tests, covid, library
Topic 4: local, time, people, arlnow, said
Topic 5: restaurant, says, said, new, open
Topic 6: said, cases, two, people, one
Topic 7: arts, community, said, covid-19, public
Topic 8: day, holiday, market, new, friday
Topic 9: cases, covid, new, reported, rate
Topic 10: said, can, says, just, community
Topic 11: said, board, million, budget, year
Topic 12: vaccine, health, covid-19, cases, booster
Topic 13: business, businesses, small, market, grant
Topic 14: can, park, covid-19, also, new
Topic 15: public, health, new, berry, care
Topic 16: people, said, one, transit, use
Topic 17: said, church, community, bus, time
Topic 18: school, aps, students, schools, said
Topic 19: said, group, office, new, people
Topic 20: barstool, freddie's, fund, bar, health
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 31 (approx. per word bound = -7.095, relative change = 1.100e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Model Converged
Beginning Spectral Initialization
Calculating the gram matrix...
Using only 10000 most frequent terms during initialization...
Finding anchor words...
..............................
Recovering initialization...
....................................................................................................
Initialization complete.
..................................................................................................................
Completed E-Step (1 seconds).
Completed M-Step.
Completing Iteration 1 (approx. per word bound = -8.003)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 2 (approx. per word bound = -7.071, relative change = 1.164e-01)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 3 (approx. per word bound = -6.946, relative change = 1.776e-02)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 4 (approx. per word bound = -6.919, relative change = 3.873e-03)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 5 (approx. per word bound = -6.909, relative change = 1.386e-03)
Topic 1: week, weekend, past, feel, anything
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, news, stories, week
Topic 5: restaurant, says, new, year, said
Topic 6: said, cases, says, dehghani-tafti, one
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, schedule
Topic 9: health, cases, reported, covid-19, said
Topic 10: can, said, just, says, community
Topic 11: school, board, said, million, budget
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, businesses, small, grant
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, transit, average, price, prices
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, said, schools
Topic 19: group, security, new, said, solar
Topic 20: health, people, help, mental, community
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, office, new
Topic 24: summer, new, said, also, school
Topic 25: employees, covid-19, mask, berry, employers
Topic 26: said, shulman, closed, people, customers
Topic 27: bus, project, pentagon, said, street
Topic 28: said, enrollment, pandemic, public, budget
Topic 29: said, work, school, time, people
Topic 30: pizza, support, pic.twitter.com, customers, #pizza
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 6 (approx. per word bound = -6.904, relative change = 6.980e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 7 (approx. per word bound = -6.901, relative change = 4.394e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 8 (approx. per word bound = -6.899, relative change = 3.019e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 9 (approx. per word bound = -6.898, relative change = 2.626e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 10 (approx. per word bound = -6.896, relative change = 2.381e-04)
Topic 1: week, weekend, said, past, feel
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, stories, news, week
Topic 5: restaurant, says, new, year, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, schedule
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, just, says, people
Topic 11: school, board, said, million, schools
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, businesses, small, grant
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, average, transit, price, prices
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, said
Topic 20: health, people, help, mental, community
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: employees, covid-19, mask, berry, employers
Topic 26: said, shulman, people, closed, customers
Topic 27: bus, project, said, pentagon, street
Topic 28: said, enrollment, pandemic, public, budget
Topic 29: said, work, time, school, people
Topic 30: support, pizza, customers, pic.twitter.com, way
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 11 (approx. per word bound = -6.894, relative change = 2.077e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 12 (approx. per word bound = -6.894, relative change = 1.392e-04)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 13 (approx. per word bound = -6.893, relative change = 9.402e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 14 (approx. per word bound = -6.892, relative change = 9.402e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 15 (approx. per word bound = -6.891, relative change = 1.101e-04)
Topic 1: week, weekend, said, past, feel
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, stories, news, people
Topic 5: restaurant, says, new, year, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, new
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, just, says, people
Topic 11: school, board, said, schools, million
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, businesses, small, can
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, average, price, prices, transit
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, immigration
Topic 20: health, people, mental, help, services
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: employees, covid-19, mask, berry, employers
Topic 26: said, shulman, people, closed, facebook
Topic 27: bus, project, said, pentagon, street
Topic 28: said, enrollment, pandemic, public, budget
Topic 29: said, work, time, people, arlnow
Topic 30: support, customers, pizza, room, tested
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 16 (approx. per word bound = -6.891, relative change = 7.633e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 17 (approx. per word bound = -6.891, relative change = 6.088e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 18 (approx. per word bound = -6.890, relative change = 4.952e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 19 (approx. per word bound = -6.890, relative change = 4.888e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 20 (approx. per word bound = -6.889, relative change = 5.718e-05)
Topic 1: week, weekend, said, past, garvey
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, news, people, week
Topic 5: restaurant, says, new, year, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, new
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, says, just, people
Topic 11: school, board, said, schools, budget
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, small, businesses, can
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, average, price, prices, transit
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, immigration
Topic 20: health, people, mental, services, help
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: employees, covid-19, mask, berry, employers
Topic 26: said, shulman, people, closed, one
Topic 27: bus, project, said, pentagon, street
Topic 28: said, enrollment, pandemic, public, budget
Topic 29: said, work, time, people, arlnow
Topic 30: support, room, customers, pizza, tested
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 21 (approx. per word bound = -6.889, relative change = 2.818e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 22 (approx. per word bound = -6.889, relative change = 1.792e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 23 (approx. per word bound = -6.889, relative change = 1.548e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 24 (approx. per word bound = -6.889, relative change = 1.455e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 25 (approx. per word bound = -6.889, relative change = 1.373e-05)
Topic 1: week, said, weekend, past, garvey
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, news, people, week
Topic 5: restaurant, says, new, year, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, new
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, says, just, people
Topic 11: school, board, said, schools, budget
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, small, businesses, can
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, average, price, prices, transit
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, immigration
Topic 20: health, people, mental, services, help
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: employees, covid-19, mask, berry, employers
Topic 26: said, shulman, people, closed, one
Topic 27: bus, project, said, pentagon, street
Topic 28: said, enrollment, pandemic, public, budget
Topic 29: said, work, time, people, arlnow
Topic 30: support, room, customers, pizza, tested
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 26 (approx. per word bound = -6.889, relative change = 1.798e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 27 (approx. per word bound = -6.889, relative change = 2.417e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 28 (approx. per word bound = -6.888, relative change = 2.675e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 29 (approx. per word bound = -6.888, relative change = 3.093e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 30 (approx. per word bound = -6.888, relative change = 4.389e-05)
Topic 1: week, said, weekend, past, garvey
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, news, people, week
Topic 5: restaurant, says, year, new, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, new
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, says, just, people
Topic 11: school, board, said, schools, budget
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, small, businesses, can
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, average, price, prices, one
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, immigration
Topic 20: health, people, mental, services, help
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: employees, covid-19, mask, berry, employers
Topic 26: said, shulman, people, closed, one
Topic 27: bus, said, project, pentagon, street
Topic 28: said, enrollment, pandemic, public, budget
Topic 29: said, work, time, people, arlnow
Topic 30: support, room, customers, pizza, tested
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 31 (approx. per word bound = -6.888, relative change = 4.872e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 32 (approx. per word bound = -6.887, relative change = 2.914e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 33 (approx. per word bound = -6.887, relative change = 2.459e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 34 (approx. per word bound = -6.887, relative change = 3.984e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 35 (approx. per word bound = -6.887, relative change = 3.578e-05)
Topic 1: week, said, weekend, past, garvey
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, news, people, week
Topic 5: restaurant, says, year, new, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, new
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, says, just, people
Topic 11: school, board, said, schools, budget
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, small, businesses, can
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, transit, average, price, prices
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, immigration
Topic 20: health, people, mental, services, help
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: employees, covid-19, mask, berry, employers
Topic 26: said, shulman, people, closed, one
Topic 27: bus, said, project, pentagon, street
Topic 28: said, enrollment, public, pandemic, budget
Topic 29: said, work, time, people, arlnow
Topic 30: tested, support, green, customers, room
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 36 (approx. per word bound = -6.886, relative change = 2.263e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 37 (approx. per word bound = -6.886, relative change = 2.625e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 38 (approx. per word bound = -6.886, relative change = 3.293e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 39 (approx. per word bound = -6.886, relative change = 2.151e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 40 (approx. per word bound = -6.886, relative change = 2.202e-05)
Topic 1: week, said, weekend, past, garvey
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, people, week, restaurants
Topic 5: restaurant, says, year, new, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, new
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, says, just, people
Topic 11: school, board, said, schools, budget
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, small, businesses, can
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, transit, average, price, prices
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, immigration
Topic 20: health, people, mental, help, services
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: employees, covid-19, mask, berry, employers
Topic 26: said, shulman, people, closed, one
Topic 27: said, bus, project, pentagon, street
Topic 28: said, enrollment, public, pandemic, budget
Topic 29: said, work, time, people, arlnow
Topic 30: tested, support, person, green, room
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 41 (approx. per word bound = -6.886, relative change = 1.722e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 42 (approx. per word bound = -6.886, relative change = 1.753e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 43 (approx. per word bound = -6.885, relative change = 1.840e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 44 (approx. per word bound = -6.885, relative change = 1.789e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 45 (approx. per word bound = -6.885, relative change = 1.894e-05)
Topic 1: week, said, weekend, past, garvey
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, people, week, restaurants
Topic 5: restaurant, says, year, new, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, new
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, says, just, people
Topic 11: school, board, said, schools, budget
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, small, businesses, can
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, transit, average, price, prices
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, immigration
Topic 20: health, people, mental, help, services
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: employees, covid-19, mask, berry, employers
Topic 26: said, shulman, people, closed, one
Topic 27: said, bus, project, pentagon, street
Topic 28: said, enrollment, public, pandemic, budget
Topic 29: said, work, time, people, arlnow
Topic 30: tested, support, person, green, said
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 46 (approx. per word bound = -6.885, relative change = 2.054e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 47 (approx. per word bound = -6.885, relative change = 1.533e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 48 (approx. per word bound = -6.885, relative change = 1.605e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 49 (approx. per word bound = -6.885, relative change = 2.252e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 50 (approx. per word bound = -6.884, relative change = 3.179e-05)
Topic 1: week, said, weekend, past, garvey
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, people, week, restaurants
Topic 5: restaurant, says, year, new, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, new
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, says, just, people
Topic 11: school, board, said, schools, budget
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, small, businesses, can
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, transit, average, price, prices
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, immigration
Topic 20: health, people, mental, help, services
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: covid-19, employees, mask, berry, employers
Topic 26: said, shulman, people, closed, one
Topic 27: said, bus, project, pentagon, street
Topic 28: said, enrollment, public, pandemic, budget
Topic 29: said, work, time, people, arlnow
Topic 30: tested, support, person, restaurants, green
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 51 (approx. per word bound = -6.884, relative change = 3.404e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 52 (approx. per word bound = -6.884, relative change = 3.901e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 53 (approx. per word bound = -6.884, relative change = 2.898e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 54 (approx. per word bound = -6.884, relative change = 2.415e-05)
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Completing Iteration 55 (approx. per word bound = -6.883, relative change = 1.432e-05)
Topic 1: week, said, weekend, past, garvey
Topic 2: said, new, arlnow, community, members
Topic 3: test, library, covid, tests, surgery
Topic 4: local, time, people, week, restaurants
Topic 5: restaurant, says, year, new, testing
Topic 6: said, cases, says, dehghani-tafti, parents
Topic 7: arts, covid-19, said, community, pandemic
Topic 8: holiday, day, closed, friday, new
Topic 9: health, cases, covid-19, reported, said
Topic 10: can, said, says, just, people
Topic 11: school, board, said, schools, budget
Topic 12: vaccine, booster, health, covid, covid-19
Topic 13: business, market, small, businesses, can
Topic 14: can, new, covid-19, also, event
Topic 15: health, monkeypox, public, care, new
Topic 16: people, transit, average, price, prices
Topic 17: church, community, said, fish, time
Topic 18: school, aps, students, schools, said
Topic 19: group, new, security, law, immigration
Topic 20: health, people, mental, help, services
Topic 21: cases, covid, new, rate, reported
Topic 22: board, said, election, theo, ballston
Topic 23: park, said, outdoor, new, office
Topic 24: summer, said, new, also, mites
Topic 25: covid-19, employees, mask, berry, employers
Topic 26: said, shulman, people, closed, one
Topic 27: said, bus, project, pentagon, street
Topic 28: said, enrollment, public, pandemic, budget
Topic 29: said, work, time, people, arlnow
Topic 30: tested, support, person, restaurants, green
..................................................................................................................
Completed E-Step (0 seconds).
Completed M-Step.
Model Converged
Code
plot(differentKs)
Code
# k=7 is looking good based on higher held out likelihood and semantic coherence and lower residuals - since it is a trade off there is no one best k, but 7 seems to have a good balance here# just using a slightly different method herearlnow_stm <-convert(arlnow_covid_dfm_new, to ="stm")diffK <-searchK(arlnow_stm$documents, arlnow_stm$vocab, K =c(3, 7, 10), N =200, data = arlnow_covid, max.em.its =1000, init.type ="Spectral")
Length Class Mode
results 8 data.frame list
call 8 -none- call
Review/More Notes
I liked working with topic models, I think it will be good to use and incorporate for answering my research questions and in my final poster. I thought the topics generally fit well, and I would like to explore them further a little bit, specifically any visualizations and further meaning. I plan to elaborate and explain further about applying this to my project. I started with supervised learning, and while it is not included here (as I didn’t find it as relevant for my data and research questions), I thought it was useful to learn about and work with. As I mentioned at the start, I plan to make my next blog post about reviewing what I have done, what I still need to/plan to do, and other loose ends as a stepping stone towards my final poster.
Source Code
---title: "Blog Post 5 (Arlnow covid articles)"author: "Miranda Manka"desription: "Working with data - Unsupervised Learning"date: "12/07/2022"format: html: toc: true code-fold: true code-copy: true code-tools: truecategories: - Miranda Manka---```{r}#| label: setup#| warning: false#| message: falselibrary(tidyverse)library(quanteda)library(quanteda.textplots)library(ggplot2)library(devtools)library(quanteda.dictionaries)library(quanteda.sentiment)library(quanteda.textmodels)library(text2vec)library(stm)library(LDAvis)knitr::opts_chunk$set(echo =TRUE, warning =FALSE, message =FALSE)```## NotesFor this blog post, I will explore some of the unsupervised learning methods from recent weeks (I worked on supervised learning methods for a bit but then realized it wasn't the best strategy for my data and what I am trying to do). At this point, I plan to make the next (and last) blog post 6 a stepping stone to my final project by examining what I have done so far and revisiting my research questions to review and make sure I have answered them.## DataThe dataset contains 457 different articles from Arlnow (local news site in Northern Virginia) from March 2020 to September 2022 (550 before removing "Morning Notes"). The code below comes from other blog posts so I will not explain it much, but it involves cleaning and pre-processing and generally getting the data ready to use.```{r}arlnow_covid =read_csv("_data/arlnow_covid_posts.csv", col_names =TRUE, show_col_types =FALSE)arlnow_covid =subset(arlnow_covid, select =-c(1))arlnow_covid = dplyr::rename(arlnow_covid, text_field = raw_text)arlnow_covid = arlnow_covid[!grepl("Morning Notes", arlnow_covid$header_text), ]arlnow_covid_corpus =corpus(arlnow_covid, docid_field ="doc_id", text_field ="text_field")arlnow_covid_summary =summary(arlnow_covid_corpus)arlnow_covid_corpus_tokens =tokens(arlnow_covid_corpus, remove_punct = T)arlnow_covid_dfm =tokens(arlnow_covid_corpus,remove_punct =TRUE, remove_symbols =TRUE,remove_numbers =TRUE) %>%dfm(tolower =TRUE) %>%dfm_remove(stopwords('english')) %>%dfm_remove(c("arlington", "county", "virginia", "$"))# most frequent terms (features)topfeatures(arlnow_covid_dfm, 20)```## Analysis - Unsupervised Learning (topic modeling)Topic models are used to identify the topics that characterize a set of documents. LDA and STM are both mixed-membership models, meaning documents are characterized as arising from a distribution over topics, rather than coming from a single topic.#### Latent Dirichlet AllocationLDA uses two basic principles: each document is made up of topics, and each word in a document can be attributed to a topic.```{r}# create new LDA modellda_model <- LDA$new(n_topics =10, doc_topic_prior =0.1,topic_word_prior =0.01)# print other methods for LDAlda_model# fitting modeldoc_topic_distr <- lda_model$fit_transform(x = arlnow_covid_dfm, n_iter =1000,convergence_tol =0.001, n_check_convergence =25,progressbar =FALSE)# doc_topic_distr is a matrix where each row is a document, each row is a topic, and the cell entry is the proportion of the document estimated to be of the topic - each row is the topic attention distribution for a document# topic distribution for the first documentbarplot(doc_topic_distr[1, ], xlab ="topic",ylab ="proportion", ylim =c(0,1),names.arg =1:ncol(doc_topic_distr))# get top n words for all topicslda_model$get_top_words(n =5, lambda =0.2)# quite a few of these topics actually seem to make sense and fit together - I think this was fairly successful and creates some interesting ideas and topics## visualization (opens in browser)#lda_model$plot()```### Structural Topic ModelsSTM allows for leveraging the information (metadata) as part of the estimation of the topics. In this case, estimating topical prevalence by incorporating time/date information in estimating the topics. ```{r}# correlated topic model - a structural topic model that incorporates covariates; an STM without covariates reduces to a very fast implementation of correlated topic models (a version of the vanilla LDA model but where the topic proportions can be positively correlated with one another).# creating a new variable from time_tag that will just be the year, then recreating the dfmtime_year =as.factor(substr(arlnow_covid$time_tag, 1, 4))arlnow_covid$time_year = time_yeararlnow_covid = arlnow_covid[-c(3)]is.character(arlnow_covid$time_year)arlnow_covid_corpus_new =corpus(arlnow_covid, docid_field ="doc_id", text_field ="text_field")arlnow_covid_dfm_new =tokens(arlnow_covid_corpus_new,remove_punct =TRUE, remove_symbols =TRUE,remove_numbers =TRUE) %>%dfm(tolower =TRUE) %>%dfm_remove(stopwords('english')) %>%dfm_remove(c("arlington", "county", "virginia", "$"))# estimate the correlated topic modelcor_topic_model <-stm(arlnow_covid_dfm_new, K =7,verbose =FALSE, init.type ="Spectral")cor_topic_model# look at the topics - the words that are most frequent, probable, or that otherwise characterize a topicsummary(cor_topic_model)# extract those wordslabelTopics(cor_topic_model)# top document associated with each topicfindThoughts(cor_topic_model,texts = arlnow_covid$text_field,topics =c(1:7),n =1)# structural topic model# I will leverage the time_tag variable included in the dataset as a covariate in my estimates of topical prevalence, I expect some topics to be more prevalent by different dates (expect topics to change over time)# specify model I chose k=7 using the searchK function, shown below/later in this documentmyModel <-stm(arlnow_covid_dfm_new,K =7,prevalence =~ time_year,data = arlnow_covid,max.em.its =1000,seed =1234,init.type ="Spectral")# view modelmyModellabelTopics(myModel)# plotting out the top topics (as groups of words associated with that topic) and their estimated frequency across the corpusplot(myModel, type ="summary")# these topics seem to make sense compared to eachother and lda and what I might expect# specify model again but without prevalence = ~ time_yearmyModel2 <-stm(arlnow_covid_dfm_new,K =7,data = arlnow_covid,max.em.its =1000,seed =1234,init.type ="Spectral")# view modelmyModel2labelTopics(myModel2)# plotting out the top topics (as groups of words associated with that topic) and their estimated frequency across the corpusplot(myModel2, type ="summary")# these topics also seem to make sense, generally# choosing k - searchK() lets you estimate a series of different models, then you can plot a series of different evaluation metrics across those choicesdifferentKs <-searchK(arlnow_covid_dfm_new,K =c(3, 5, 7, 10, 20, 30),prevalence =~ time_year,N =200,data = arlnow_covid,max.em.its =1000,init.type ="Spectral")plot(differentKs)# k=7 is looking good based on higher held out likelihood and semantic coherence and lower residuals - since it is a trade off there is no one best k, but 7 seems to have a good balance here# just using a slightly different method herearlnow_stm <-convert(arlnow_covid_dfm_new, to ="stm")diffK <-searchK(arlnow_stm$documents, arlnow_stm$vocab, K =c(3, 7, 10), N =200, data = arlnow_covid, max.em.its =1000, init.type ="Spectral")plot(diffK)summary(diffK)```## Review/More NotesI liked working with topic models, I think it will be good to use and incorporate for answering my research questions and in my final poster. I thought the topics generally fit well, and I would like to explore them further a little bit, specifically any visualizations and further meaning. I plan to elaborate and explain further about applying this to my project.I started with supervised learning, and while it is not included here (as I didn't find it as relevant for my data and research questions), I thought it was useful to learn about and work with. As I mentioned at the start, I plan to make my next blog post about reviewing what I have done, what I still need to/plan to do, and other loose ends as a stepping stone towards my final poster.