library(dplyr)
contributing_df_filepath <-"/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/metadata/CONTRIBUTING_weekly_count_data.csv"
contributing_df = read.csv(contributing_df_filepath, header = TRUE) 

readme_df_filepath <- "/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/metadata/README_weekly_count_data.csv"
readme_df = read.csv(readme_df_filepath, header = TRUE)

combined_df <- bind_rows(
  contributing_df %>%
    group_by(project_id) %>%
    select(project_id, age_at_commit) %>%
    mutate(document = factor("CONTRIBUTING", levels = c("CONTRIBUTING", "README"))),
  readme_df %>%
    group_by(project_id) %>%
    select(project_id, age_at_commit) %>%
    mutate(document = factor("README", levels = c("CONTRIBUTING", "README")))
)

unique_combined_df <- combined_df %>%
  distinct(project_id, age_at_commit, document)

library(tidyverse)
library(tidyquant)
library(ggdist)
library(ggthemes)
library(ggplot2)

age_raincloud <- unique_combined_df |>
  ggplot(aes(x = factor(document), y = age_at_commit, fill = factor(document))) +
  geom_boxplot(
    width = 0.12,
    # removing outliers
    outlier.color = NA,
    alpha = 0.5
  ) +
  ggplot::stat_dots(
    # ploting on left side
    side = "left",
    # adjusting position
    justification = 1.1,
    # adjust grouping (binning) of observations
    binwidth = 0.25
  )

age_raincloud