Tasks for Week 05

Published

September 19, 2024

UK Election Data

library(tidyverse)

── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr     1.1.4     ✔ readr     2.1.5
✔ forcats   1.0.0     ✔ stringr   1.5.1
✔ ggplot2   3.5.1     ✔ tibble    3.2.1
✔ lubridate 1.9.3     ✔ tidyr     1.3.1
✔ purrr     1.0.2     
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

1. Install the UK Election Data package

It’s not on CRAN, it’s on my GitHub.

# You only need to do this once
remotes::install_github("kjhealy/ukelection2019")

2. Load the package

library(ukelection2019)

ukvote2019

# A tibble: 3,320 × 13
   cid     constituency electorate party_name candidate votes vote_share_percent
   <chr>   <chr>             <int> <chr>      <chr>     <int>              <dbl>
 1 W07000… Aberavon          50747 Labour     Stephen … 17008               53.8
 2 W07000… Aberavon          50747 Conservat… Charlott…  6518               20.6
 3 W07000… Aberavon          50747 The Brexi… Glenda D…  3108                9.8
 4 W07000… Aberavon          50747 Plaid Cym… Nigel Hu…  2711                8.6
 5 W07000… Aberavon          50747 Liberal D… Sheila K…  1072                3.4
 6 W07000… Aberavon          50747 Independe… Captain …   731                2.3
 7 W07000… Aberavon          50747 Green      Giorgia …   450                1.4
 8 W07000… Aberconwy         44699 Conservat… Robin Mi… 14687               46.1
 9 W07000… Aberconwy         44699 Labour     Emily Ow… 12653               39.7
10 W07000… Aberconwy         44699 Plaid Cym… Lisa Goo…  2704                8.5
# ℹ 3,310 more rows
# ℹ 6 more variables: vote_share_change <dbl>, total_votes_cast <int>,
#   vrank <int>, turnout <dbl>, fname <chr>, lname <chr>

Each row is a candidate standing in a particular constituency (in US speak, a district) for a particular party or as an independent candidate.

3. Get familiar with the data

Use sample_n() to sample n rows of your tibble.

ukvote2019 |> 
  sample_n(10)

# A tibble: 10 × 13
   cid     constituency electorate party_name candidate votes vote_share_percent
   <chr>   <chr>             <int> <chr>      <chr>     <int>              <dbl>
 1 S14000… Banff & Buc…      66655 Conservat… David Du… 21182               50.1
 2 W07000… Brecon & Ra…      55490 Monster R… Lady Lil…   345                0.8
 3 E14000… Thirsk & Ma…      80979 Independe… Steve Mu…   245                0.4
 4 E14001… Wealden           83038 Conservat… Nus Ghani 37043               60.8
 5 E14000… Hampshire N…      83083 Conservat… Kit Malt… 36591               62.1
 6 E14000… Worthing Ea…      75195 Liberal D… Ashley R…  4127                7.8
 7 E14000… Sefton Cent…      69760 Green      Alison G…  1261                2.5
 8 E14000… Barnsley Ea…      69504 Liberal D… Sophie T…  1330                3.5
 9 E14000… Leeds North…      70580 Green      Rachel H…  1931                3.8
10 E14000… Sunderland …      72672 Independe… Dale McK…   484                1.1
# ℹ 6 more variables: vote_share_change <dbl>, total_votes_cast <int>,
#   vrank <int>, turnout <dbl>, fname <chr>, lname <chr>

A vector of unique constituency names:

ukvote2019 |> 
  distinct(constituency)

# A tibble: 650 × 1
   constituency                   
   <chr>                          
 1 Aberavon                       
 2 Aberconwy                      
 3 Aberdeen North                 
 4 Aberdeen South                 
 5 Aberdeenshire West & Kincardine
 6 Airdrie & Shotts               
 7 Aldershot                      
 8 Aldridge-Brownhills            
 9 Altrincham & Sale West         
10 Alyn & Deeside                 
# ℹ 640 more rows

Tally them up:

ukvote2019 |> 
  distinct(constituency) |> 
  tally()

# A tibble: 1 × 1
      n
  <int>
1   650

That is, there are 650 electoral constituencies in Great Britain and Northern Ireland.

A quicker way of establishing how many constituencies there are:

ukvote2019 |> 
  count(constituency)

# A tibble: 650 × 2
   constituency                        n
   <chr>                           <int>
 1 Aberavon                            7
 2 Aberconwy                           4
 3 Aberdeen North                      6
 4 Aberdeen South                      4
 5 Aberdeenshire West & Kincardine     4
 6 Airdrie & Shotts                    5
 7 Aldershot                           4
 8 Aldridge-Brownhills                 5
 9 Altrincham & Sale West              6
10 Alyn & Deeside                      5
# ℹ 640 more rows

Which parties fielded the most candidates?

ukvote2019 |> 
  count(party_name) |> 
  arrange(desc(n))

# A tibble: 69 × 2
   party_name                     n
   <chr>                      <int>
 1 Conservative                 636
 2 Labour                       631
 3 Liberal Democrat             611
 4 Green                        497
 5 The Brexit Party             275
 6 Independent                  224
 7 Scottish National Party       59
 8 UKIP                          44
 9 Plaid Cymru                   36
10 Christian Peoples Alliance    29
# ℹ 59 more rows

What are the Top 5 parties by n candidates?

ukvote2019 |> 
  count(party_name) |> 
  slice_max(order_by = n, n = 5)

# A tibble: 5 × 2
  party_name           n
  <chr>            <int>
1 Conservative       636
2 Labour             631
3 Liberal Democrat   611
4 Green              497
5 The Brexit Party   275

Bottom 5? Does this make sense?

ukvote2019 |> 
  count(party_name) |> 
  slice_min(order_by = n, n = 5)

# A tibble: 25 × 2
   party_name                              n
   <chr>                               <int>
 1 Ashfield Independents                   1
 2 Best for Luton                          1
 3 Birkenhead Social Justice Party         1
 4 British National Party                  1
 5 Burnley & Padiham Independent Party     1
 6 Church of the Militant Elvis Party      1
 7 Citizens Movement Party UK              1
 8 CumbriaFirst                            1
 9 Heavy Woollen District Independents     1
10 Independent Network                     1
# ℹ 15 more rows

4. Filtering

Filtering is subsetting the rows according to a condition in one or more of the columns

Show me all and only the Green party candidates.

ukvote2019 |> 
  filter(party_name == "Green")

# A tibble: 497 × 13
   cid     constituency electorate party_name candidate votes vote_share_percent
   <chr>   <chr>             <int> <chr>      <chr>     <int>              <dbl>
 1 W07000… Aberavon          50747 Green      Giorgia …   450                1.4
 2 S14000… Aberdeen No…      62489 Green      Guy Inge…   880                2.4
 3 S14000… Airdrie & S…      64008 Green      Rosemary…   685                1.7
 4 E14000… Aldershot         72617 Green      Donna Wa…  1750                3.7
 5 E14000… Aldridge-Br…      60138 Green      Bill McC…   771                2  
 6 E14000… Altrincham …      73096 Green      Geraldin…  1566                2.9
 7 E14000… Amber Valley      69976 Green      Lian Piz…  1388                3  
 8 E14000… Arundel & S…      81726 Green      Isabel T…  2519                4.1
 9 E14000… Ashfield          78204 Green      Rose Woo…   674                1.4
10 E14000… Ashford           89550 Green      Mandy Ro…  2638                4.4
# ℹ 487 more rows
# ℹ 6 more variables: vote_share_change <dbl>, total_votes_cast <int>,
#   vrank <int>, turnout <dbl>, fname <chr>, lname <chr>

Show me all candidates named “Michael”.

ukvote2019 |> 
  filter(fname == "Michael")

# A tibble: 25 × 13
   cid     constituency electorate party_name candidate votes vote_share_percent
   <chr>   <chr>             <int> <chr>      <chr>     <int>              <dbl>
 1 E14000… Basildon So…      74441 Liberal D… Michael …  1957                4.3
 2 N06000… Belfast Sou…      69984 Ulster Un… Michael …  1259                2.7
 3 E14000… Blaydon           67853 The Brexi… Michael …  5833               12.8
 4 E14000… Bosworth          81537 Liberal D… Michael …  9096               16.1
 5 E14000… Bury South        75152 Independe… Michael …   277                0.6
 6 E14000… Canterbury        80203 Independe… Michael …   505                0.8
 7 W07000… Cardiff Nor…      68438 Green      Michael …   820                1.6
 8 E14000… Dorset Mid …      65426 Conservat… Michael … 29548               60.4
 9 S14000… Dundee East       66210 Liberal D… Michael …  3573                7.9
10 E14000… Durham Nort…      72166 Liberal D… Michael …  2831                5.9
# ℹ 15 more rows
# ℹ 6 more variables: vote_share_change <dbl>, total_votes_cast <int>,
#   vrank <int>, turnout <dbl>, fname <chr>, lname <chr>

Show me all Green party candidates named “Michael”.

ukvote2019 |> 
  filter(party_name == "Green" & fname == "Michael")

# A tibble: 3 × 13
  cid      constituency electorate party_name candidate votes vote_share_percent
  <chr>    <chr>             <int> <chr>      <chr>     <int>              <dbl>
1 W070000… Cardiff Nor…      68438 Green      Michael …   820                1.6
2 E140007… Gloucester        81332 Green      Michael …  1385                2.6
3 E140008… Preston           59672 Green      Michael …   660                2  
# ℹ 6 more variables: vote_share_change <dbl>, total_votes_cast <int>,
#   vrank <int>, turnout <dbl>, fname <chr>, lname <chr>

5. Grouping

Who won in each constituency?

ukvote2019 |> 
  group_by(constituency) |> 
  slice_max(votes)

# A tibble: 650 × 13
# Groups:   constituency [650]
   cid     constituency electorate party_name candidate votes vote_share_percent
   <chr>   <chr>             <int> <chr>      <chr>     <int>              <dbl>
 1 W07000… Aberavon          50747 Labour     Stephen … 17008               53.8
 2 W07000… Aberconwy         44699 Conservat… Robin Mi… 14687               46.1
 3 S14000… Aberdeen No…      62489 Scottish … Kirsty B… 20205               54  
 4 S14000… Aberdeen So…      65719 Scottish … Stephen … 20388               44.7
 5 S14000… Aberdeenshi…      72640 Conservat… Andrew B… 22752               42.7
 6 S14000… Airdrie & S…      64008 Scottish … Neil Gray 17929               45.1
 7 E14000… Aldershot         72617 Conservat… Leo Doch… 27980               58.4
 8 E14000… Aldridge-Br…      60138 Conservat… Wendy Mo… 27850               70.8
 9 E14000… Altrincham …      73096 Conservat… Graham B… 26311               48  
10 W07000… Alyn & Dees…      62783 Labour     Mark Tami 18271               42.5
# ℹ 640 more rows
# ℹ 6 more variables: vote_share_change <dbl>, total_votes_cast <int>,
#   vrank <int>, turnout <dbl>, fname <chr>, lname <chr>

What happens if you leave out group_by() in the chunk of code above?

How do I count the number of seats each party won?

ukvote2019 |> 
  group_by(constituency) |> 
  slice_max(votes) |> 
  group_by(party_name) |> 
  tally() |> 
  arrange(desc(n))

# A tibble: 10 × 2
   party_name                           n
   <chr>                            <int>
 1 Conservative                       366
 2 Labour                             202
 3 Scottish National Party             48
 4 Liberal Democrat                    11
 5 Democratic Unionist Party            8
 6 Sinn Féin                            7
 7 Plaid Cymru                          4
 8 Social Democratic & Labour Party     2
 9 Alliance Party                       1
10 Green                                1

Group and Summarize

ukvote2019 |> 
  group_by(constituency) |> 
  slice_max(votes) |> 
  ungroup() |> 
  summarize(mean_winner_share = mean(vote_share_percent))

# A tibble: 1 × 1
  mean_winner_share
              <dbl>
1              54.4

What happens if you leave out ungroup() in the chunk above?

4. Have a go

Can you find …

The candidate who won the most votes in the country?
The candidate with the largest vote share in the country?
The median vote share of winning candidates?
The largest vote share swing from previous election?
Overall turnout for the whole country?
Median turnout across constituencies?