adding matter for updating CR

mlm/contributing_crescendo_model.R

@@ -0,0 +1,24 @@
+library(dplyr)
+library(lubridate)
+library(rdd)
+
+contributing_df_filepath <- "/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/CONTRIBUTING_weekly_count_data.csv"
+contributing_df = read.csv(contributing_df_filepath, header = TRUE) 
+
+window_num <- 5
+contributing_df <- contributing_df |>
+  filter(week_index >= (- window_num) & week_index <= (window_num)) |>
+  mutate(scaled_age = scale(age)) |>
+  mutate(scaled_age_at_commit = scale(age_at_commit))|>
+  mutate(log1p_count = log1p(commit_count)) |>
+  mutate(C = ifelse(week_index > -4 & week_index <= 0, 1,0))
+
+library(lme4)
+library(optimx)
+library(lattice)
+
+all_gmodel <- glmer.nb(log1p_count ~ C * week_index + scaled_age + scaled_age_at_commit + (C * week_index | project_id),
+                       control=glmerControl(optimizer="bobyqa",
+                                            optCtrl=list(maxfun=2e5)), nAGQ=0, data=contributing_df)
+summary(all_gmodel)
+saveRDS(all_gmodel, "020425_contributing_crescendo_model.rda")

quartile_splits.R

@@ -0,0 +1,80 @@
+contributing_count_filepath <- "/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/CONTRIBUTING_weekly_count_data.csv"
+contributing_count_df = read.csv(contributing_count_filepath, header = TRUE) 
+
+window_num <- 5
+contributing_count_df <- contributing_count_df |>
+  filter(week_index >= (- window_num) & week_index <= (window_num)) |>
+  mutate(scaled_age = scale(age)) |>
+  mutate(scaled_age_at_commit = scale(age_at_commit))|>
+  mutate(log1p_count = log1p(commit_count))
+
+quantile(contributing_count_df$age / 365, probs = c(0.25, 0.5, 0.75))
+quantile(contributing_count_df$age_at_commit / 365, probs = c(0.25, 0.5, 0.75))
+aggregate_cccd <- contributing_count_df |>
+  group_by(project_id) |>
+  summarize(
+    avg_weekly_commits = mean(commit_count),
+    authors_before = sum(new_author_emails[before_after == 0]),
+    authors_after = sum(new_author_emails[before_after == 1]),
+  )
+quantile(aggregate_cccd$avg_weekly_commits, probs = c(0.25, 0.5, 0.75))
+quantile(aggregate_cccd$authors_before, probs = c(0.25, 0.5, 0.75))
+quantile(aggregate_cccd$authors_after, probs = c(0.25, 0.5, 0.75))
+
+readme_count_filepath <- "/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/README_weekly_count_data.csv"
+readme_count_df = read.csv(readme_count_filepath, header = TRUE) 
+
+window_num <- 5
+readme_count_df <- readme_count_df |>
+  filter(week_index >= (- window_num) & week_index <= (window_num)) |>
+  mutate(scaled_age = scale(age)) |>
+  mutate(scaled_age_at_commit = scale(age_at_commit))|>
+  mutate(log1p_count = log1p(commit_count))
+
+quantile(readme_count_df$age / 365, probs = c(0.25, 0.5, 0.75))
+quantile(readme_count_df$age_at_commit / 365, probs = c(0.25, 0.5, 0.75))
+aggregate_rccd <- readme_count_df |>
+  group_by(project_id) |>
+  summarize(
+    avg_weekly_commits = mean(commit_count),
+    authors_before = sum(new_author_emails[before_after == 0]),
+    authors_after = sum(new_author_emails[before_after == 1]),
+  )
+quantile(aggregate_rccd$avg_weekly_commits, probs = c(0.25, 0.5, 0.75))
+quantile(aggregate_rccd$authors_before, probs = c(0.25, 0.5, 0.75))
+quantile(aggregate_rccd$authors_after, probs = c(0.25, 0.5, 0.75))
+
+readme_readability <- read.csv("text_analysis/020325_README_readability.csv", header=TRUE)
+median(readme_readability$mcalpine_eflaw)
+median(readme_readability$linsear_write_formula)
+median(readme_readability$flesch_reading_ease)
+
+contributing_readability <- read.csv("text_analysis/020125_CONTRIBUTING_readability.csv", header=TRUE)
+median(contributing_readability$mcalpine_eflaw)
+median(contributing_readability$linsear_write_formula)
+median(contributing_readability$flesch_reading_ease)
+
+
+contributing_topic <- read.csv("text_analysis/020125_CONTRIBUTING_file_topic_distributions.csv", header=TRUE)
+quantile(contributing_topic$t0, probs = c(0.25, 0.5, 0.75))
+quantile(contributing_topic$t1, probs = c(0.25, 0.5, 0.75))
+quantile(contributing_topic$t2, probs = c(0.25, 0.5, 0.75))
+quantile(contributing_topic$t3, probs = c(0.25, 0.5, 0.75))
+quantile(contributing_topic$t4, probs = c(0.25, 0.5, 0.75))
+
+readme_topic <- read.csv("text_analysis/020325_README_file_topic_distributions.csv", header=TRUE)
+quantile(readme_topic$t0, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t1, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t2, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t3, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t4, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t5, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t6, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t7, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t8, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t9, probs = c(0.25, 0.5, 0.75))
+quantile(readme_topic$t10, probs = c(0.25, 0.5, 0.75))
+
+
+
+

text_analysis/checking_overlap.R

@@ -0,0 +1,4 @@
+contributing_manifest <- read.csv("text_analysis/0203_contributing_merged_manifest.csv", header=TRUE)
+readme_manifest <- read.csv("text_analysis/0203_readme_merged_manifest.csv", header=TRUE)
+
+overlap = inner_join(contributing_manifest, readme_manifest, by="repo_id")

topic-outcome-models/topic_model_presentation.R

@@ -0,0 +1,14 @@
+library(texreg)
+
+readme_rdd <- readRDS("topic-outcome-models/020325_README_commit_topic_model.rda")
+contrib_rdd <- readRDS("topic-outcome-models/020325_CONTRIBUTING_commit_topic_model.rda")
+
+texreg(readme_rdd, stars=NULL, digits=3, use.packages=FALSE, 
+       custom.model.names=c( 'README'),
+       custom.coef.names=c('Topic 1', 'Topic 2', 'Topic 3', 'Topic 4', 'Topic 5', 'Topic 6', 'Topic 7', 'Topic 8', 'Topic 9', 'Topic 10', 'Topic 11'), 
+       table=FALSE, ci.force = TRUE)
+
+texreg(contrib_rdd, stars=NULL, digits=3, use.packages=FALSE, 
+       custom.model.names=c( 'CONTRIBUITING'),
+       custom.coef.names=c('Topic 1', 'Topic 2', 'Topic 3', 'Topic 4', 'Topic 5'), 
+       table=FALSE, ci.force = TRUE)