Using dplyr to count multiple group-by variables

Question

I have a data-set with multiple categorical variables

data <- data_frame(
HomeTeam = c("Team1", "Team2", "Team3", "Team4", "Team2", "Team2", "Team4", 
             "Team3", "Team2", "Team1", "Team3", "Team2"),
AwayTeam = c("Team2", "Team1", "Team4", "Team3", "Team1", "Team4", "Team1", 
             "Team2", "Team3", "Team3", "Team4", "Team1"),
HomeScore = c(10, 5, 12, 18, 17, 19, 23, 17, 34, 19, 8, 3),
AwayScore = c(4, 16, 9, 19, 16, 4, 8, 21, 6, 5, 9, 17),
Venue = c("Ground1", "Ground2", "Ground3", "Ground3", "Ground1", "Ground2", 
          "Ground1", "Ground3", "Ground2", "Ground3", "Ground4", "Ground2"))

I basically want to summarise "HomeTeam" and "AwayTeam" by count into a new table, as per below

 HomeTeam NumberOfGamesHome NumberOfGamesaWAY
 <chr>                <int>             <int>
 1 Team1                    2                 4
 2 Team2                    5                 2
 3 Team3                    3                 3
 4 Team4                    2                 3

My current approach requires two group-by lines of code, then joining the tables

HomeTeamCount <- data %>% 
group_by(HomeTeam) %>% 
summarise(NumberOfGamesHome = n()) 

AwayTeamCount <- data %>% 
group_by(AwayTeam) %>% 
summarise(NumberOfGamesAway = n()) 

Desired <- left_join(HomeTeamCount, AwayTeamCount, 
                 by = c("HomeTeam" = "AwayTeam"))

In my actual data-set, I have a large number of categorical variables, and following the above approach seems laborious and inefficient

Is there a way with dplyr to group_by multiple categorical variables, to produce the desired output? Or potentially data.table?

I have consulted several other questions such as here and here, but have not been able to figure out the answer.

Maurits Evers Maurits Evers · Accepted Answer · 2018-11-24T06:14:49

Here is one possibility using gather to spread data from wide to long, grouping by teams and summarising the number of home and away games.

library(tidyverse)
data %>%
    gather(key, Team) %>%
    group_by(Team) %>%
    summarise(
        NumberOfGamesHome = sum(key == "HomeTeam"),
        NumberOfGamesaWAY = sum(key == "AwayTeam"))
## A tibble: 4 x 3
#  Team  NumberOfGamesHome NumberOfGamesaWAY
#  <chr>             <int>             <int>
#1 Team1                 2                 4
#2 Team2                 5                 2
#3 Team3                 3                 3
#4 Team4                 2                 3

Update

To reflect your updated sample data with additional columns you can do

data %>%
    gather(key, Team, HomeTeam, AwayTeam) %>%
    group_by(Team) %>%
    summarise(
        NumberOfGamesHome = sum(key == "HomeTeam"),
        NumberOfGamesaWAY = sum(key == "AwayTeam"))
## A tibble: 4 x 3
#  Team  NumberOfGamesHome NumberOfGamesaWAY
#  <chr>             <int>             <int>
#1 Team1                 2                 4
#2 Team2                 5                 2
#3 Team3                 3                 3
#4 Team4                 2                 3

Using dplyr to count multiple group-by variables

1 Answers

Update