Code
library(tidyverse)
library(dplyr)
library(readr)
library(readxl)
knitr::opts_chunk$set(echo = TRUE, warning=FALSE, message=FALSE)Challenge 4 Post
challenge_4
abc_poll
More data wrangling: pivoting
Challenge Overview
Today’s challenge is to:
- read in a data set, and describe the data set using both words and any supporting information (e.g., tables, etc)
- tidy data (as needed, including sanity checks)
- identify variables that need to be mutated
- mutate variables and sanity check all mutations
Read in data
Read in one (or more) of the following datasets, using the correct R package and command.
- For challenge 4, I will be reading in abc_poll.csv ⭐
Code
abc_poll <- read.csv("_data/abc_poll_2021.csv")
abc_pollCode
dim(abc_poll)[1] 527 31Code
summary(abc_poll) id xspanish complete_status ppage
Min. :7230001 Length:527 Length:527 Min. :18.00
1st Qu.:7230132 Class :character Class :character 1st Qu.:40.00
Median :7230264 Mode :character Mode :character Median :55.00
Mean :7230264 Mean :53.39
3rd Qu.:7230396 3rd Qu.:67.00
Max. :7230527 Max. :91.00
ppeduc5 ppeducat ppgender ppethm
Length:527 Length:527 Length:527 Length:527
Class :character Class :character Class :character Class :character
Mode :character Mode :character Mode :character Mode :character
pphhsize ppinc7 ppmarit5 ppmsacat
Length:527 Length:527 Length:527 Length:527
Class :character Class :character Class :character Class :character
Mode :character Mode :character Mode :character Mode :character
ppreg4 pprent ppstaten PPWORKA
Length:527 Length:527 Length:527 Length:527
Class :character Class :character Class :character Class :character
Mode :character Mode :character Mode :character Mode :character
ppemploy Q1_a Q1_b Q1_c
Length:527 Length:527 Length:527 Length:527
Class :character Class :character Class :character Class :character
Mode :character Mode :character Mode :character Mode :character
Q1_d Q1_e Q1_f Q2
Length:527 Length:527 Length:527 Length:527
Class :character Class :character Class :character Class :character
Mode :character Mode :character Mode :character Mode :character
Q3 Q4 Q5 QPID
Length:527 Length:527 Length:527 Length:527
Class :character Class :character Class :character Class :character
Mode :character Mode :character Mode :character Mode :character
ABCAGE Contact weights_pid
Length:527 Length:527 Min. :0.3240
Class :character Class :character 1st Qu.:0.6332
Mode :character Mode :character Median :0.8451
Mean :1.0000
3rd Qu.:1.1516
Max. :6.2553 Code
n_distinct(abc_poll$ppstaten)[1] 49Briefly describe the data
The ABC Poll dataset is most likely capturing responses from a US-based nationwide poll. Running the n_distinct () function for the state or ppstaten variable determines that responses from 49 US states were featured The dim () function reveals that there are 527 rows or observations and 31 columns or variables. We can infer from the row count that 527 people responded to the poll. From the summary () function, we can deduce that the majority of the variables seem to have character values, such as education level, gender, marital status, age, region, employment status, etc–aimed at collecting demographic data and political data. There are other variables, such as the ID number variable, the Q1-Q5 variables, the Contact variable that serve to help better organize the information and respondent preferences for further interviews collected by the poll.
Tidy Data (as needed)
Is your data already tidy, or is there work to be done? Be sure to anticipate your end result to provide a sanity check, and document your work here.
The dataset is already tidy in the sense that every value does have its own cell, each variable has its own column, and each observation has its own row.
Identify variables that need to be mutated
Are there any variables that require mutation to be usable in your analysis stream? For example, are all time variables correctly coded as dates? Are all string variables reduced and cleaned to sensible categories? Do you need to turn any variables into factors and reorder for ease of graphics and visualization?
In terms of variables that require mutation, in this dataset, I mainly believe this takes shape in the form of reducing and cleaning some of the string variables to more sensible categories.
Mutating the QPID variable
In the QPID or political party identification variable, we can take the “A or An” out of cells so a response within this column, like “An Independent” becomes “Independent.”
Running the table () function shows that in addition to included political party responses of “A Democrat,” “A Republican,” and “An Independent,” there are also two types of responses in the QPID column: “Skipped” or “Something Else.” The “Something Else” response can remain as is but we can re-code the “Skipped” value to read as NA instead of “Skipped.”
Code
# view the QPID column within the dataset
table(abc_poll$QPID)
A Democrat A Republican An Independent Skipped Something else
176 152 168 3 28 Code
# mutate so that "A" or "An" can be removed from party identification responses and so that "Skipped" can read as NA
abc_poll2 <- abc_poll%>%
mutate(party_id = str_remove(QPID, "A[n]*"),
party_id = case_when(
str_detect(QPID, "Skipped")~NA_character_,
TRUE~party_id
)) %>%
select(-QPID)
abc_poll2Code
# check
table(abc_poll2$party_id)
Democrat Independent Republican Something else
176 168 152 28 Code
unique(abc_poll2$party_id)[1] " Democrat" " Independent" "Something else" " Republican"
[5] NA Seems QPID variable has been renamed party_id. As expected, it also seems like “Skipped” has been changed to NA and “Democrat,” “Republican” and “Independent” responses no longer have the word “A” or “An” before them!
Changing “Skipped” to NA from in responses within the Q1-Q5 columns/variables
When poll participants didn’t offer a response when answering the Q1-Q5 questions in the poll, a response of “Skipped” was recorded. Like we did in the previous section, we can transform this into reading as NA, for analysis purposes. Instead of mutating one ‘Q’ column at a time, we can do them simultaneously by using the across() function within mutate ().
Code
# mutating "Skipped" to NA in the Q1-Q5 columns/variables
abc_poll2<-abc_poll2%>%
mutate(across(starts_with("Q"), ~ na_if(.x, "Skipped")))
abc_poll2Code
#checking one of the 'Q' columns, Q1_b, to see if the NA value has taken the place of "Skipped"
table(abc_poll2$Q1_b)
Approve Disapprove
192 322 Code
unique(abc_poll2$Q1_b)[1] "Approve" "Disapprove" NA Looks like NA has taken the place of “Skipped” in Q1_b (and assumingly all the Q columns) as expected!