The group_by, summarize, and ungroup verbs

Data Manipulation with dplyr

James Chapman

Curriculum Manager, DataCamp

Summarize

counties %>%
  summarize(total_population = sum(population))

# A tibble: 1 x 1
  total_population
             <dbl>
1        315845353

Aggregate and summarize

counties %>%
  summarize(total_population = sum(population),
            average_unemployment = mean(unemployment))

# A tibble: 1 x 2
  total_population average_unemployment
             <dbl>                <dbl>
1        315845353                 7.80

Summary functions

sum()
mean()
median()
min()
max()
n()

Aggregate within groups

counties %>%
  group_by(state) %>%

  summarize(total_pop = sum(population),
            average_unemployment = mean(unemployment))

# A tibble: 50 x 3
   state       total_pop average_unemployment
   <chr>           <dbl>                <dbl>
 1 Alabama       4830620                758. 
 2 Alaska         725461                257. 
 3 Arizona       6641928                180. 
 4 Arkansas      2958208                674. 
 5 California   38421464                626. 
 6 Colorado      5278906                477. 
 7 Connecticut   3593222                 65.3
 8 Delaware       926454                 23.8
 9 Florida      19645772                696. 
10 Georgia      10006693               1586. 
# … with 40 more rows

Sorting summaries

counties %>%
  group_by(state) %>%
  summarize(total_pop = sum(population),
            average_unemployment = mean(unemployment)) %>%

  arrange(desc(average_unemployment))

# A tibble: 50 x 3
   state          total_pop average_unemployment
   <chr>              <dbl>                <dbl>
 1 Mississippi      2988081                12.0 
 2 Arizona          6641928                12.0 
 3 South Carolina   4777576                11.3 
 4 Alabama          4830620                11.3 
 5 California      38421464                10.8 
 6 Nevada           2798636                10.5 
 7 North Carolina   9845333                10.5 
 8 Florida         19645772                10.4 
 9 Georgia         10006693                 9.97
10 Michigan         9900571                 9.96
# … with 40 more rows

Metro column

counties %>%
  select(state, metro, county, population)

# A tibble: 3,138 x 4
   state   metro    county   population
   <chr>   <chr>    <chr>         <dbl>
 1 Alabama Metro    Autauga       55221
 2 Alabama Metro    Baldwin      195121
 3 Alabama Nonmetro Barbour       26932
 4 Alabama Metro    Bibb          22604
 5 Alabama Metro    Blount        57710
 6 Alabama Nonmetro Bullock       10678
 7 Alabama Nonmetro Butler        20354
 8 Alabama Metro    Calhoun      116648
 9 Alabama Nonmetro Chambers      34079
10 Alabama Nonmetro Cherokee      26008
# … with 3,128 more rows

Grouping on multiple columns

counties %>%
  group_by(state, metro) %>%
  summarize(total_pop = sum(population))

# A tibble: 97 x 3
# Groups:   state [50]
   state      metro    total_pop
   <chr>      <chr>        <dbl>
 1 Alabama    Metro      3671377
 2 Alabama    Nonmetro   1159243
 3 Alaska     Metro       494990
 4 Alaska     Nonmetro    230471
 5 Arizona    Metro      6295145
 6 Arizona    Nonmetro    346783
 7 Arkansas   Metro      1806867
 8 Arkansas   Nonmetro   1151341
 9 California Metro     37587429
10 California Nonmetro    834035
# … with 87 more rows

Ungroup

counties %>%
  group_by(state, metro) %>%
  summarize(total_pop = sum(population)) %>%
  ungroup()

# A tibble: 97 x 3
   state      metro    total_pop
   <chr>      <chr>        <dbl>
 1 Alabama    Metro      3671377
 2 Alabama    Nonmetro   1159243
 3 Alaska     Metro       494990
 4 Alaska     Nonmetro    230471
 5 Arizona    Metro      6295145
 6 Arizona    Nonmetro    346783
 7 Arkansas   Metro      1806867
 8 Arkansas   Nonmetro   1151341
 9 California Metro     37587429
10 California Nonmetro    834035
# … with 87 more rows

Let's practice!

Data Manipulation with dplyr