Creating cohort tables from event data is complicated and requires several lines of code. The cohorts package lets users convert data frames to cohort tables in both long and wide formats with simple functions. Users may choose between day and month level cohorts.
You can install the released version of cohorts from CRAN with:
install.packages("cohorts")And the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("PeerChristensen/cohorts")In this example, we use a dataset consisting of customer IDs and invoice dates.
library(cohorts)
head(online_cohorts)
#> CustomerID InvoiceDate
#> 1 17850 2010-12-01
#> 2 13047 2010-12-01
#> 3 12583 2010-12-01
#> 4 13748 2010-12-01
#> 5 15100 2010-12-01
#> 6 15291 2010-12-01We can then turn this into a cohort table where each customer ID is tracked from the first invoice month until the last month in the period.
online_cohorts %>%
cohort_table_month(CustomerID, InvoiceDate)
#> # A tibble: 13 × 14
#> cohort `Dec 2010` `Jan 2011` `Feb 2011` `Mar 2011` `Apr 2011` `May 2011`
#> <int> <int> <int> <int> <int> <int> <int>
#> 1 1 949 363 318 368 342 377
#> 2 2 NA 421 101 119 102 138
#> 3 3 NA NA 380 94 73 106
#> 4 4 NA NA NA 440 84 112
#> 5 5 NA NA NA NA 299 68
#> 6 6 NA NA NA NA NA 279
#> 7 7 NA NA NA NA NA NA
#> 8 8 NA NA NA NA NA NA
#> 9 9 NA NA NA NA NA NA
#> 10 10 NA NA NA NA NA NA
#> 11 11 NA NA NA NA NA NA
#> 12 12 NA NA NA NA NA NA
#> 13 13 NA NA NA NA NA NA
#> # … with 7 more variables: `Jun 2011` <int>, `Jul 2011` <int>,
#> # `Aug 2011` <int>, `Sep 2011` <int>, `Oct 2011` <int>, `Nov 2011` <int>,
#> # `Dec 2011` <int>If we need to track activity on a daily basis, we can instead use the
cohort_table_day() function.
gamelaunch %>%
cohort_table_day(userid, eventDate)
#> # A tibble: 31 × 32
#> cohort `2016-04-27` `2016-04-28` `2016-04-29` `2016-04-30` `2016-05-01`
#> <int> <int> <int> <int> <int> <int>
#> 1 1 96 65 55 46 46
#> 2 2 NA 200 117 96 84
#> 3 3 NA NA 370 207 181
#> 4 4 NA NA NA 387 223
#> 5 5 NA NA NA NA 405
#> 6 6 NA NA NA NA NA
#> 7 7 NA NA NA NA NA
#> 8 8 NA NA NA NA NA
#> 9 9 NA NA NA NA NA
#> 10 10 NA NA NA NA NA
#> # … with 21 more rows, and 26 more variables: `2016-05-02` <int>,
#> # `2016-05-03` <int>, `2016-05-04` <int>, `2016-05-05` <int>,
#> # `2016-05-06` <int>, `2016-05-07` <int>, `2016-05-08` <int>,
#> # `2016-05-09` <int>, `2016-05-10` <int>, `2016-05-11` <int>,
#> # `2016-05-12` <int>, `2016-05-13` <int>, `2016-05-14` <int>,
#> # `2016-05-15` <int>, `2016-05-16` <int>, `2016-05-17` <int>,
#> # `2016-05-18` <int>, `2016-05-19` <int>, `2016-05-20` <int>, …In order to see the percent of remaining customers in subsequent
periods, we can pipe the above code into the cohort_table_pct()
function.
gamelaunch %>%
cohort_table_day(userid, eventDate) %>%
cohort_table_pct(decimals = 1)
#> # A tibble: 31 × 32
#> cohort `2016-04-27` `2016-04-28` `2016-04-29` `2016-04-30` `2016-05-01`
#> <int> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 1 100 67.7 57.3 47.9 47.9
#> 2 2 NA 100 58.5 48 42
#> 3 3 NA NA 100 55.9 48.9
#> 4 4 NA NA NA 100 57.6
#> 5 5 NA NA NA NA 100
#> 6 6 NA NA NA NA NA
#> 7 7 NA NA NA NA NA
#> 8 8 NA NA NA NA NA
#> 9 9 NA NA NA NA NA
#> 10 10 NA NA NA NA NA
#> # … with 21 more rows, and 26 more variables: `2016-05-02` <dbl>,
#> # `2016-05-03` <dbl>, `2016-05-04` <dbl>, `2016-05-05` <dbl>,
#> # `2016-05-06` <dbl>, `2016-05-07` <dbl>, `2016-05-08` <dbl>,
#> # `2016-05-09` <dbl>, `2016-05-10` <dbl>, `2016-05-11` <dbl>,
#> # `2016-05-12` <dbl>, `2016-05-13` <dbl>, `2016-05-14` <dbl>,
#> # `2016-05-15` <dbl>, `2016-05-16` <dbl>, `2016-05-17` <dbl>,
#> # `2016-05-18` <dbl>, `2016-05-19` <dbl>, `2016-05-20` <dbl>, …Another option is to shift cohort tables left. Here, we align cohorts such that date columns are replaced by time periods, i.e. t0, t1, t2 etc.
To left-shift a cohort table, we can use the shift_left() function.
gamelaunch %>%
cohort_table_day(userid, eventDate) %>%
shift_left()
#> # A tibble: 31 × 32
#> cohort t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10
#> <int> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 1 96 65 55 46 46 45 44 33 34 31 26
#> 2 2 200 117 96 84 82 76 62 72 63 52 59
#> 3 3 370 207 181 152 138 127 114 98 95 89 84
#> 4 4 387 223 177 151 129 122 107 115 114 86 88
#> 5 5 405 222 178 152 130 131 128 103 86 98 84
#> 6 6 325 183 146 125 119 105 85 73 72 59 59
#> 7 7 270 165 129 113 113 102 85 89 75 74 72
#> 8 8 264 142 124 91 73 76 81 63 60 55 55
#> 9 9 267 153 114 110 99 94 89 72 68 62 65
#> 10 10 127 74 58 51 42 42 50 41 42 40 32
#> # … with 21 more rows, and 20 more variables: t11 <dbl>, t12 <dbl>, t13 <dbl>,
#> # t14 <dbl>, t15 <dbl>, t16 <dbl>, t17 <dbl>, t18 <dbl>, t19 <dbl>,
#> # t20 <dbl>, t21 <dbl>, t22 <dbl>, t23 <dbl>, t24 <dbl>, t25 <dbl>,
#> # t26 <dbl>, t27 <dbl>, t28 <dbl>, t29 <dbl>, t30 <dbl>We can also get the raw numbers as percentages.
gamelaunch %>%
cohort_table_day(userid, eventDate) %>%
shift_left_pct()
#> # A tibble: 31 × 32
#> cohort t0 t1 t2 t3 t4 t5 t6 t7 t8 t9 t10
#> <int> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 1 100 67.7 57.3 47.9 47.9 46.9 45.8 34.4 35.4 32.3 27.1
#> 2 2 100 58.5 48 42 41 38 31 36 31.5 26 29.5
#> 3 3 100 55.9 48.9 41.1 37.3 34.3 30.8 26.5 25.7 24.1 22.7
#> 4 4 100 57.6 45.7 39 33.3 31.5 27.6 29.7 29.5 22.2 22.7
#> 5 5 100 54.8 44 37.5 32.1 32.3 31.6 25.4 21.2 24.2 20.7
#> 6 6 100 56.3 44.9 38.5 36.6 32.3 26.2 22.5 22.2 18.2 18.2
#> 7 7 100 61.1 47.8 41.9 41.9 37.8 31.5 33 27.8 27.4 26.7
#> 8 8 100 53.8 47 34.5 27.7 28.8 30.7 23.9 22.7 20.8 20.8
#> 9 9 100 57.3 42.7 41.2 37.1 35.2 33.3 27 25.5 23.2 24.3
#> 10 10 100 58.3 45.7 40.2 33.1 33.1 39.4 32.3 33.1 31.5 25.2
#> # … with 21 more rows, and 20 more variables: t11 <dbl>, t12 <dbl>, t13 <dbl>,
#> # t14 <dbl>, t15 <dbl>, t16 <dbl>, t17 <dbl>, t18 <dbl>, t19 <dbl>,
#> # t20 <dbl>, t21 <dbl>, t22 <dbl>, t23 <dbl>, t24 <dbl>, t25 <dbl>,
#> # t26 <dbl>, t27 <dbl>, t28 <dbl>, t29 <dbl>, t30 <dbl>To visualize the data, we can turn a cohort table into long format and create a line plot.
In this example, we select only the first seven cohorts.
library(tidyverse)
gamelaunch_long <- gamelaunch %>%
cohort_table_day(userid, eventDate) %>%
shift_left_pct() %>%
pivot_longer(-cohort) %>%
mutate(time = as.numeric(str_remove(name,"t")))
gamelaunch_long %>%
filter(value > 0, cohort <= 7, time > 0) %>%
ggplot(aes(time, value, colour = factor(cohort), group = cohort)) +
geom_line(size = 1.5) +
geom_point(size = 1.5) +
theme_light()
Another way to plot a cohort table is by means of tiles. In this case we provide the percentages and colour the tiles accordingly.
gamelaunch_long %>%
filter(time > 0, value > 0) %>%
ggplot(aes(time, reorder(cohort, desc(cohort)))) +
geom_raster(aes(fill = log(value))) +
coord_equal(ratio = 1) +
geom_text(aes(label = glue::glue("{round(value,0)}%")), size = 2, color = "snow") +
scale_fill_gradient(guide = F) +
theme_minimal() +
theme(panel.grid = element_blank(),
panel.border = element_blank()) +
labs(y= "cohort")