In team-oss/scrape-cran: Measuring the Cost and Value of Open Source Software

# run this instead of clicking `knit`
rmarkdown::render(here::here('./src/exploratory/code_gov/_____.Rmd'),
                  output_dir = here::here('./output/code_gov'))
bad_html <- './src/exploratory/code_gov/_____.html'
if (file.exists(here::here(bad_html))) file.remove(here::here(bad_html))

knitr::opts_chunk$set(echo = TRUE)
library(magrittr)
library(here)
library(tibble)
library(dplyr)
library(urltools)

source(here::here('./R/file_paths.R'))

code_gov_df <- readr::read_delim(here::here(.GlobalEnv$CODE_GOV_CLEANED_FILE), delim = '\t')

addmargins(table(code_gov_df$agency.acronym,
                 code_gov_df$status,
                 useNA = 'always'))

sub_cols <- code_gov_df %>%
  select(-contains('tags'), -contains('languages'), -contains('permission'), -vcs, -starts_with('X.'))
names(sub_cols)

hist(code_gov_df$laborHours[code_gov_df$laborHours > -1 & code_gov_df$laborHours < 10])

hist(code_gov_df$laborHours[code_gov_df$laborHours > 10 & code_gov_df$laborHours < 1000])

hist(code_gov_df$laborHours[code_gov_df$laborHours > 10 & code_gov_df$laborHours < 100])

hist(code_gov_df$score)

my_mode <- function(x) {
  uni <- unique(x)
  uni[which.max(tabulate(match(x, uni)))]
}

code_gov_df %>%
  dplyr::group_by(agency.acronym) %>%
  dplyr::summarize(n = n(),
                   missing_labor_values = sum(is.na(laborHours)),
                   num_labor_values = sum(!is.na(laborHours)),
                   labor_stat_mean = mean(laborHours, na.rm = TRUE),
                   labor_stat_median = median(laborHours, na.rm = TRUE),
                   labor_stat_mode = my_mode(laborHours)
                   ) %>%
  dplyr::arrange(n) %>%
  print(n = Inf)

library(ggplot2)

ggplot(data = code_gov_df) + geom_histogram(aes(x = laborHours)) + ggplot2::scale_x_log10()

team-oss/scrape-cran documentation built on Dec. 23, 2021, 8:42 a.m.