Challenge 2

Data wrangling: using group() and summarise()

Setup

library(tidyverse)

knitr::opts_chunk$set(echo = TRUE, warning=FALSE, message=FALSE)

Read in the Data

For this challenge I will be using the birds.csv dataset.

Code

birds <- read_csv("_data/birds.csv")
head(birds)

# A tibble: 6 × 14
  `Domain Code` Domain      `Area Code` Area  `Element Code` Element `Item Code`
  <chr>         <chr>             <dbl> <chr>          <dbl> <chr>         <dbl>
1 QA            Live Anima…           2 Afgh…           5112 Stocks         1057
2 QA            Live Anima…           2 Afgh…           5112 Stocks         1057
3 QA            Live Anima…           2 Afgh…           5112 Stocks         1057
4 QA            Live Anima…           2 Afgh…           5112 Stocks         1057
5 QA            Live Anima…           2 Afgh…           5112 Stocks         1057
6 QA            Live Anima…           2 Afgh…           5112 Stocks         1057
# ℹ 7 more variables: Item <chr>, `Year Code` <dbl>, Year <dbl>, Unit <chr>,
#   Value <dbl>, Flag <chr>, `Flag Description` <chr>

Lets print out the Column names to check which columns we are interested in.

Code

colnames(birds)

 [1] "Domain Code"      "Domain"           "Area Code"        "Area"            
 [5] "Element Code"     "Element"          "Item Code"        "Item"            
 [9] "Year Code"        "Year"             "Unit"             "Value"           
[13] "Flag"             "Flag Description"

All the Columns that contain “Code” in them can be removed as we have the normal columns as well. For example, since we have the “Area” column, we can remove the “Area Code” column.

Now, lets go through the unique values of certain columns to see if we need them.

Code

unique(birds$`Flag Description`)

[1] "FAO estimate"                                                                
[2] "Official data"                                                               
[3] "FAO data based on imputation methodology"                                    
[4] "Data not available"                                                          
[5] "Unofficial figure"                                                           
[6] "Aggregate, may include official, semi-official, estimated or calculated data"

The “Flag Description” and “Flag” columns give us information about the manner of data collection and the authenticity of the data and can be removed for our analysis.

Code

unique(birds$Unit)

[1] "1000 Head"

Code

unique(birds$Domain)

[1] "Live Animals"

The “Unit” and “Domain” columns contain only one unique value and hence can be removed.

Now we will take a subset of the data with only the columns that we require and perform our analysis.

Code

birds_subset <- birds[c( "Area", "Element", "Item", "Year", "Value")]
head(birds_subset)

# A tibble: 6 × 5
  Area        Element Item      Year Value
  <chr>       <chr>   <chr>    <dbl> <dbl>
1 Afghanistan Stocks  Chickens  1961  4700
2 Afghanistan Stocks  Chickens  1962  4900
3 Afghanistan Stocks  Chickens  1963  5000
4 Afghanistan Stocks  Chickens  1964  5300
5 Afghanistan Stocks  Chickens  1965  5500
6 Afghanistan Stocks  Chickens  1966  5800

Code

birds_subset <- na.omit(birds_subset)

Describe the data

By looking at the data we can infer that it is about the stock of various poultry birds for major Areas in the world, for each year. Lets take an example of a row in the data set:

Code

birds_subset[1,]

# A tibble: 1 × 5
  Area        Element Item      Year Value
  <chr>       <chr>   <chr>    <dbl> <dbl>
1 Afghanistan Stocks  Chickens  1961  4700

This shows that, for the year 1961, the stock of poultry Chickens in Afghanistan was 4700.

Code

summary(birds_subset)

     Area             Element              Item                Year     
 Length:29941       Length:29941       Length:29941       Min.   :1961  
 Class :character   Class :character   Class :character   1st Qu.:1976  
 Mode  :character   Mode  :character   Mode  :character   Median :1992  
                                                          Mean   :1991  
                                                          3rd Qu.:2005  
                                                          Max.   :2018  
     Value         
 Min.   :       0  
 1st Qu.:     171  
 Median :    1800  
 Mean   :   99411  
 3rd Qu.:   15404  
 Max.   :23707134  

From the above summary we see that the range of years in this data set goes from 1961 to 2018.

We will now check the unique values for the other columns: Areas:

Code

length(unique(birds_subset$Area))

[1] 248

Items:

Code

unique(birds_subset$Item)

[1] "Chickens"               "Ducks"                  "Geese and guinea fowls"
[4] "Turkeys"                "Pigeons, other birds"  

Grouped Summary Statistics

Analysis of Items

We will first group the data by Items to see which Poultry Item has been more popular. We will sort it by descending order

Code

birds_subset%>%
  group_by(Item)%>%
  summarise(Avg_stocks = mean(Value))%>%
  arrange(desc(Avg_stocks))

# A tibble: 5 × 2
  Item                   Avg_stocks
  <chr>                       <dbl>
1 Chickens                  207931.
2 Ducks                      23072.
3 Turkeys                    15228.
4 Geese and guinea fowls     10292.
5 Pigeons, other birds        6163.

We can see that Chickens have been the most popular Poultry Item across the world as its stock is significantly higher than the other items. The sum of the stocks of all other items is also less than the stock for Chickens.

Analysis of Areas

Code

birds_subset%>%
  filter(Year==2010)%>%
  group_by(Area)%>%
  summarise(Total_stocks = sum(Value))%>%
  arrange(desc(Total_stocks))

# A tibble: 238 × 2
   Area               Total_stocks
   <chr>                     <dbl>
 1 World                  22311214
 2 Asia                   12535841
 3 Eastern Asia            6906604
 4 China, mainland         6312000
 5 Americas                5635098
 6 South-eastern Asia      2570583
 7 Northern America        2379199
 8 South America           2299071
 9 Europe                  2265162
10 Southern Asia           2222019
# ℹ 228 more rows

Here, we can see that the larger continental areas, have a larger stock. This makes sense as the aggregate population of a continent will be greater than that of a country and will have higher consumption.

Asia has the highest poultry consumption. This could be due to the high population density in Asia which adds to the higher consumption.

United States of America is the country with the highest poultry consumption, while Falkland Islands (Malvinas) has the lowest consumption.

Analysis by decade

We can see the trends of poultry as we go through various decades.

Code

birds_subset %>%
  mutate(Decade = 10 * floor(Year / 10)) %>%
  group_by(Item,Decade) %>%
  summarise(Avg_stocks = mean(Value),
            Median_stocks = median(Value),
            Standard_Deviation = sd(Value))

# A tibble: 30 × 5
# Groups:   Item [5]
   Item     Decade Avg_stocks Median_stocks Standard_Deviation
   <chr>     <dbl>      <dbl>         <dbl>              <dbl>
 1 Chickens   1960     83286.         4745             360089.
 2 Chickens   1970    111927.         7000             481532.
 3 Chickens   1980    162130.         9200             703217.
 4 Chickens   1990    216889.        12000            1007193.
 5 Chickens   2000    287001.        15375            1375191.
 6 Chickens   2010    361437.        18719            1732596.
 7 Ducks      1960      8323.          372.             31291.
 8 Ducks      1970     11006.          430.             42341.
 9 Ducks      1980     16187.          560.             63333.
10 Ducks      1990     26600.          485             113131.
# ℹ 20 more rows

Here we observe that for each Item, the consumption has gradually increased over the decades. The increase in standard deviation tells us that a few countries are consuming a large chunk of the poultry in the World