updating some plots for results section, also saving model to file

dsl/dsl.R

@@ -1,7 +1,7 @@
 library(tidyverse)
 library(dsl)
 
-dsl_csv <-"111725_DSL_frame.csv"
+dsl_csv <-"~/dsl/111725_DSL_frame.csv"
 dsl_df <- read.csv(dsl_csv, header = TRUE) 
 
 dsl_df <- dsl_df |>
@@ -81,7 +81,7 @@ dev_model <- dsl(
   data=dsl_df
 )
 summary(dev_model)
-
+saveRDS(dev_model, "120225_logit_dsl.RDS")
 
 library(broom)
 library(dplyr)
@@ -101,6 +101,43 @@ tidy.dsl <- function(x, conf.int = FALSE, conf.level = 0.95, exponentiate = FALS
   return(out)
 }
 coef_df <- tidy.dsl(dev_model)
+coef_df <- coef_df |>
+  mutate(
+    term = recode(term,
+                  "week_index" = "Weeks from deployment",
+                  "(Intercept)" = "Intercept",
+                  "n_comments_before" = "# of comments prior to resolution",
+                  "median_PC4_adac" = "Median Author PC4 Pre-resolution",
+                  "median_PC3_adac" = "Median Author PC3 Pre-resolution",
+                  "median_gerrit_reviewers" = "Median # of Code Reviewers (Gerrit)",
+                  "median_gerrit_loc_delta" = "Median LoC Changed (Gerrit)",
+                  "human_TSOL_prop_adac" = "% of sentences discussing 'Solutions'",
+                  "human_RK_prop_adac" = "% of sentences discussing 'Record Keeping'",
+                  "human_EP_prop_adac" = "% of sentences discussing 'Existent Problems'",
+                  "as.factor(source)c3" = "HTTP-deprecation (factor)",
+                  "as.factor(source)c2" = "HTTPS-as-default (factor)",
+                  "as.factor(isAuthorWMF)TRUE" = "WMF-affiliate Author (factor)",
+                  "as.factor(isAuthorWMF)TRUE:as.factor(source)c2" = "WMF-affiliate Author:HTTPS-as-default",
+                  "as.factor(isAuthorWMF)TRUE:as.factor(source)c3" = "WMF-affiliate Author:HTTP-deprecation",
+                  ),
+    term = factor(term, levels = rev(c(
+      "Intercept",   
+      "% of sentences discussing 'Existent Problems'",        
+      "% of sentences discussing 'Solutions'",
+      "% of sentences discussing 'Record Keeping'",
+      "Median Author PC4 Pre-resolution",
+      "Median Author PC3 Pre-resolution",
+      "# of comments prior to resolution",
+      "Median # of Code Reviewers (Gerrit)",
+      "Median LoC Changed (Gerrit)",
+      "Weeks from deployment",
+      "HTTPS-as-default (factor)",
+      "HTTP-deprecation (factor)",
+      "WMF-affiliate Author (factor)",
+      "WMF-affiliate Author:HTTPS-as-default",
+      "WMF-affiliate Author:HTTP-deprecation"
+    )))
+  )
 ggplot(coef_df, aes(x = estimate, y = term)) +
   geom_point(size = 1) +
   geom_errorbar(aes(xmin = estimate - 1.96*std.error, xmax = estimate + 1.96 *std.error), height = 0.2) +

dsl/final_bivariate.R

@@ -93,7 +93,8 @@ ggplot(
   geom_point() +
   geom_smooth() + 
   scale_color_viridis_d() +
-  theme_minimal() 
+  theme_minimal() + 
+  labs(x = "Weeks from Release", y = "% of sentences machine-tagged as'Existent Problems'", title = "Proportion of 'Existent Problems' tags over time")
 
 dsl_df <- dsl_df |>
   mutate(priority = factor(priority, 

p2/quest/neurobiber_PCA_analysis.R

@@ -12,6 +12,92 @@ library(dplyr)
 main_csv <- "~/analysis_data/110925_unified.csv"
 main_df <- read.csv(main_csv  , header = TRUE)
 
+main_df |>
+  ggplot(
+    aes(
+      x = PC4,
+      y = PC3,
+      fill = comment_type
+    )
+  ) + 
+  facet_grid(~source, scales="fixed",
+             labeller = as_labeller(c(
+               "c1" = "VisualEditor (c1)",
+               "c2"   = "HTTPS-as-default (c2)",
+               "c3"  = "HTTP-deprecation (c3)"
+             ))) +
+  geom_point(shape = 21, alpha=0.3, size=2) +
+  xlim(-50, 50) + 
+  ylim(-50, 50) +
+  scale_fill_viridis_d(
+    option = "magma",
+    name = "Comment type",
+    labels = c("Task Description", "Reply"))+
+  theme_minimal() +
+  theme(legend.position = "top") +
+  labs(
+    title = "PCs for Task Comments by comment type and case",
+    x = "Casual v. Formal Updates (PC3)",
+    y = "Technical-matter v. Procedural Commentary (PC4)",
+  )
+
+main_df |>
+  filter(ADAC=="1") |>
+  ggplot(
+    aes(
+      x = PC4,
+      y = PC3,
+      fill = isAuthorWMF
+    )
+  ) + 
+  facet_grid(comment_type~source,
+             labeller = as_labeller(c(
+               "c1" = "VisualEditor (c1)",
+               "c2"   = "HTTPS-as-default (c2)",
+               "c3"  = "HTTP-deprecation (c3)",
+               "task_description" = "Task Description",
+               "task_subcomment" = "Follow-up Reply"
+             ))) +
+  geom_point(shape = 21, alpha=0.3, size=2) +
+  scale_fill_viridis_d(
+    name = "Comment Author Affiliation",
+    labels = c("Nonaffiliated", "WMF-affiliated"))+
+  theme_minimal() +
+  theme(legend.position = "top") +
+  labs(
+    title = "PCs for Pre-Resolution Comments Written by Task Author (by Author Affiliation, Case, and Comment Type)",
+    x = "Casual v. Formal Updates (PC3)",
+    y = "Technical-matter v. Procedural Commentary (PC4)",
+  )
+
+main_df |>
+  filter(comment_type=="task_subcomment") |>
+  ggplot(
+    aes(
+      x = PC4,
+      y = PC3,
+      fill = as.factor(ADAC)
+    )
+  ) + 
+  facet_grid(isAuthorWMF~source,
+             labeller = as_labeller(c(
+               "c1" = "VisualEditor (c1)",
+               "c2"   = "HTTPS-as-default (c2)",
+               "c3"  = "HTTP-deprecation (c3)"
+             ))) +
+  geom_point(shape = 21, alpha=0.13, size=2) +
+  scale_fill_viridis_d(
+    option = "turbo",
+    name = "By Task Author Before Resolution",
+    labels = c("No", "Yes"))+
+  theme_minimal() +
+  theme(legend.position = "top") +
+  labs(
+    title = "PCs for Replies (by Author Affiliation, Case, and Comment Type)",
+    x = "Casual v. Formal Updates (PC3)",
+    y = "Technical-matter v. Procedural Commentary (PC4)",
+  )
+
 main_df <- main_df |>
   mutate(
     comment_wordcount = as.integer(stringr::str_count(tidyr::replace_na(as.character(comment_text), ""), "\\S+"))