contributing_model_done

mlm/readme_did_model_fit.R

@@ -16,10 +16,10 @@ library(lme4)
 library(optimx)
 library(lattice)
 
-#all_gmodel <- glmer.nb(log1p_count ~ before_after * week_index + scaled_age + (before_after * week_index | project_id),
+all_gmodel <- glmer.nb(log1p_count ~ before_after * week_index + scaled_age + (before_after * week_index | project_id),
-#                       control=glmerControl(optimizer="bobyqa",
+                       control=glmerControl(optimizer="bobyqa",
-#                                            optCtrl=list(maxfun=2e5)), nAGQ=0,
+                                            optCtrl=list(maxfun=2e5)), nAGQ=0,
-#                       data=readme_df)
+                       data=readme_df)
 
 summary(all_gmodel)
 #saveRDS(all_gmodel, "020125_readme_model.rda")
@@ -49,5 +49,3 @@ g <- glmer_ranef_Dweek |>
 g
 
 write.csv(glmer_ranef_Dweek, "0201_readme_dweek_ranefs.csv")
-
-

mlm/readme_did_prep.R

@@ -6,6 +6,12 @@ library(rdd)
 readme_df_filepath <- "/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/README_weekly_count_data.csv"
 df = read.csv(readme_df_filepath, header = TRUE)
 
+#filtered_df <- df |>
+#  filter(!project_id == "letsencrypt_letsencrypt")
+
+#output_filepath <-"/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/README_weekly_count_data.csv"
+#write.csv(filtered_df, output_filepath, row.names = FALSE)
+
 #EDA 
 var(df$commit_count) # 112.4945
 mean(df$commit_count) # 2.431342

topic-outcome-models/contributing_topic_outcome_model.R

@@ -0,0 +1,37 @@
+library(dplyr)
+library(lubridate)
+library(rdd)
+library(stringr)
+
+contributing_count_data_filepath <- "/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/CONTRIBUTING_weekly_count_data.csv"
+contributing_count_df = read.csv(contributing_count_data_filepath, header = TRUE) 
+
+contributing_topic_dist_filepath <- "text_analysis/020125_CONTRIBUTING_file_topic_distributions.csv"
+contributing_topics_df = read.csv(contributing_topic_dist_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))
+
+summed_data <- contributing_count_df |>
+  filter(before_after == 1) |>
+  group_by(project_id) |>
+  summarise_at(vars(commit_count), list(summed_count=sum))
+
+contributing_topics_df <- contributing_topics_df |>
+  mutate(project_id = sapply(str_split(filename, "_hullabaloo_"), `[`, 1)) |>
+  mutate(project_id = ifelse(filename=="_vcr_vcr_CONTRIBUTING.md", "vcr_vcr", project_id)) |>
+  mutate(project_id = ifelse(filename=="marshmallow-code_marshmallow.git_CONTRIBUTING.rst", "marshmallow-code_marshmallow.git", project_id))
+
+merged_df <- inner_join(summed_data, contributing_topics_df, by="project_id")
+merged_df$logged_commits <- log1p(merged_df$summed_count)
+  
+library(MASS)
+
+commit_outcome_model <- glm.nb(logged_commits ~ 0 + t0 + t1 + t2 + t3 + t4, data=merged_df)
+qqnorm(residuals(commit_outcome_model))
+summary(commit_outcome_model)
+saveRDS(commit_outcome_model, "020325_commit_topic_model.rda")

topic-outcome-models/readme_topic_outcome_model.R

@@ -0,0 +1,30 @@
+library(dplyr)
+library(lubridate)
+library(rdd)
+library(stringr)
+
+readme_count_data_filepath <- "/mmfs1/gscratch/comdata/users/mjilg/govdoc-cr-data/final_data/README_weekly_count_data.csv"
+readme_count_df = read.csv(readme_count_data_filepath, header = TRUE) 
+
+readme_topic_dist_filepath <- "text_analysis/020125_README_file_topic_distributions.csv"
+readme_topics_df = read.csv(readme_topic_dist_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))
+
+summed_data <- readme_count_df |>
+  filter(before_after == 1) |>
+  group_by(project_id) |>
+  summarise_at(vars(commit_count), list(summed_count=sum))
+
+readme_topics_df <- readme_topics_df |>
+  mutate(project_id = sapply(str_split(filename, "_hullabaloo_"), `[`, 1)) |> 
+  mutate(project_id = ifelse(filename=="jaraco_keyrings.alt_hullabaloo_README.rst", "jaraco_keyrings.alt", project_id)) |>
+  mutate(project_id = ifelse(filename=="_vcr_vcr_README.md", "vcr_vcr", project_id)) 
+  
+merged_df <- inner_join(summed_data, readme_topics_df, by="project_id")
+merged_df$logged_commits <- log1p(merged_df$summed_count)