mgaughan/mw-lifecycle-analysis
1
0
Fork 0
Code

adding trial survival test and more information about adac variables

Browse Source
This commit is contained in:
Matthew Gaughan 2025-10-27 17:54:14 -07:00
parent ab1cb3efea
commit 2efd961fed
7 changed files with 3369 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

BIN
.RData
View File

Binary file not shown.

34
analysis_data/style_dict_variables.R
View File

@ -4,9 +4,39 @@ library(tidyr)
library(dplyr)
library(purrr)
unified_csv <-"~/analysis_data/102425_unified.csv"
unified_csv <-"~/analysis_data/102725_unified.csv"
unified_df <- read.csv(unified_csv, header = TRUE)
unified_df |>
ggplot(
aes(
x=leng,
y=as.factor(isAuthorWMF)
)
) +
geom_boxplot(alpha = 0.7, position = position_dodge(width = 0.9)) +
facet_grid(. ~ source, scales = "fixed") + # Facet by source; adjust as needed
scale_fill_viridis_d() +
theme_minimal()
BE_set <- c("OBSERVED BUG BEHAVIOR", "EXPECTED BEHAVIOR")
SOL_set <- c("SOLUTION DISCUSSION", "SOLUTION USAGE")
@ -168,7 +198,7 @@ ggplot(second_join, aes(x = modal_verbs, y = PC1, color=comment_type)) +
ggplot(second_join, aes(
x = as.factor(comment_type), # x-axis grouping
y = olmo_VR_prop,
y = modal_verbs,
fill = isAuthorWMF
)) +
ylim(0, 3) +

3130
dsl/102725_DSL_df_adac.csv Normal file
View File

File diff suppressed because it is too large Load Diff

64
dsl/dsl_aggregation.R
View File

@ -81,9 +81,34 @@ human_list_unified_df <- unified_df %>%
mean(list_human_labels[ADAC == 1] %in% c("BUG REPRODUCTION",
"INVESTIGATION AND EXPLORATION"), na.rm = TRUE)
),
# ADAC==0 proportions
n_tags_no_adac = sum(!is.na(list_human_labels) & ADAC == 0),
human_BE_prop_no_adac = if_else(
n_tags_no_adac == 0L,
NA_real_,
mean(list_human_labels[ADAC == 0] %in% c("OBSERVED BUG BEHAVIOR", "EXPECTED BEHAVIOR"), na.rm = TRUE)
),
human_SOL_prop_no_adac = if_else(
n_tags_no_adac == 0L,
NA_real_,
mean(list_human_labels[ADAC == 0] %in% c("SOLUTION DISCUSSION", "SOLUTION USAGE"), na.rm = TRUE)
),
human_VR_prop_no_adac = if_else(
n_tags_no_adac == 0L,
NA_real_,
mean(list_human_labels[ADAC == 0] %in% c("OBSERVED BUG BEHAVIOR", "EXPECTED BEHAVIOR",
"SOLUTION DISCUSSION", "SOLUTION USAGE",
"INVESTIGATION AND EXPLORATION"), na.rm = TRUE)
),
human_BI_prop_no_adac = if_else(
n_tags_no_adac == 0L,
NA_real_,
mean(list_human_labels[ADAC == 0] %in% c("BUG REPRODUCTION",
"INVESTIGATION AND EXPLORATION"), na.rm = TRUE)
),
.groups = "drop"
) |>
select(-n_tags, -n_tags_adac)
select(-n_tags, -n_tags_adac, -n_tags_no_adac)
olmo_list_unified_df <- unified_df %>%
@ -156,9 +181,33 @@ olmo_list_unified_df <- unified_df %>%
mean(olmo_label[ADAC == 1] %in% c("BUG REPRODUCTION",
"INVESTIGATION AND EXPLORATION"), na.rm = TRUE)
),
n_tags_no_adac = sum(!is.na(olmo_label) & ADAC == 0),
olmo_BE_prop_no_adac = if_else(
n_tags_no_adac == 0L,
NA_real_,
mean(olmo_label[ADAC == 0] %in% c("OBSERVED BUG BEHAVIOR", "EXPECTED BEHAVIOR"), na.rm = TRUE)
),
olmo_SOL_prop_no_adac = if_else(
n_tags_no_adac == 0L,
NA_real_,
mean(olmo_label[ADAC == 0] %in% c("SOLUTION DISCUSSION", "SOLUTION USAGE"), na.rm = TRUE)
),
olmo_VR_prop_no_adac = if_else(
n_tags_no_adac == 0L,
NA_real_,
mean(olmo_label[ADAC == 0] %in% c("OBSERVED BUG BEHAVIOR", "EXPECTED BEHAVIOR",
"SOLUTION DISCUSSION", "SOLUTION USAGE",
"INVESTIGATION AND EXPLORATION"), na.rm = TRUE)
),
olmo_BI_prop_no_adac = if_else(
n_tags_no_adac == 0L,
NA_real_,
mean(olmo_label[ADAC == 0] %in% c("BUG REPRODUCTION",
"INVESTIGATION AND EXPLORATION"), na.rm = TRUE)
),
.groups = "drop"
) |>
select(-n_tags, -n_tags_adac)
select(-n_tags, -n_tags_adac, -n_tags_no_adac)
# aggregate other Task-level variables and then join
task_level_variables <- unified_df |>
@ -166,7 +215,14 @@ task_level_variables <- unified_df |>
summarise(median_gerrit_loc_delta = median(gerrit_code_insertions + gerrit_code_deletions, na.rm = TRUE),
median_gerrit_reviewers = median(gerrit_reviewer_count, na.rm = TRUE),
median_PC3 = median(PC3),
median_PC3_ADAC = median(PC3[ADAC==1])
median_PC3_adac = median(PC3[ADAC==1]),
median_PC3_no_adac = median(PC3[ADAC==0]),
median_PC1 = median(PC1),
median_PC1_adac = median(PC1[ADAC==1]),
median_PC1_no_adac = median(PC1[ADAC==0]),
median_PC4 = median(PC4),
median_PC4_adac = median(PC4[ADAC==1]),
median_PC4_no_adac = median(PC4[ADAC==0]),
)
descriptions <- unified_df |>
@ -247,4 +303,4 @@ ggplot(task_level_variables, aes(
y = "Time to Resolution (up to 60 days)",
)
# 4. save
write.csv(task_level_variables, "102725_DSL_df.csv", row.names = FALSE)
write.csv(task_level_variables, "102725_DSL_df_adac.csv", row.names = FALSE)

28
dsl/survival.R Normal file
View File

@ -0,0 +1,28 @@
library(tidyverse)
dsl_csv <-"~/dsl/102725_DSL_df_adac.csv"
dsl_df <- read.csv(dsl_csv, header = TRUE)
#https://stats.oarc.ucla.edu/wp-content/uploads/2025/02/survival_r_full.html
dsl_df <- dsl_df |>
filter(source == "c1")
library(survival)
library(broom)
dsl_df$ttr_weeks <- dsl_df$TTR / 168
trial.survival <- Surv(dsl_df$ttr_weeks)
trial.model <- coxph(trial.survival ~ isAuthorWMF +
median_PC3_adac + week_index +
median_gerrit_loc_delta + median_gerrit_reviewers +
olmo_BI_prop_adac, data=dsl_df)
summary(trial.model)
trial.tab <- tidy(trial.model, exponentiate=T, conf.int=T)
ggplot(trial.tab,
aes(y=term, x=estimate, xmin=conf.low, xmax=conf.high)) +
geom_pointrange() + # plots center point (x) and range (xmin, xmax)
geom_vline(xintercept=1, color="red") + # vertical line at HR=1
labs(x="hazard ratio", title="Hazard ratios and 95% CIs") +
theme_classic()
surv.at.means <- survfit(trial.model)
plot(surv.at.means, xlab="weeks", ylab="survival probability")

41
p2/quest/adac_analysis.R Normal file
View File

@ -0,0 +1,41 @@
library(tidyverse)
main_csv <- "~/dsl/102725_DSL_df_adac.csv"
main_df <- read.csv(main_csv , header = TRUE)
main_df <- main_df |>
mutate(
pc_adac_delta = median_PC4_no_adac - median_PC4_adac,
olmo_BI_adac_delta = olmo_BI_prop_no_adac - olmo_BI_prop_adac
)
ggplot(main_df, aes(
x = as.factor(phase), # x-axis grouping
y = olmo_BI_adac_delta,
fill = resolution_outcome
)) +
ylim(-3, 3) +
geom_boxplot(alpha = 0.7, position = position_dodge(width = 0.9)) +
facet_grid(. ~ source, scales = "fixed") + # Facet by source; adjust as needed
scale_fill_viridis_d() +
theme_minimal() +
labs(
title = "Boxplot of PC4",
x = "Comment_type",
y = "PC4",
fill = "isAuthorWMF?"
)
ggplot(main_df, aes(x = week_index,
y = median_PC3_adac, fill = resolution_outcome)) +
facet_grid(~source, scales="fixed") +
geom_point(shape = 21, alpha=0.3, size=2) +
scale_fill_viridis_d() +
theme_minimal() +
labs(
title = "PCs for Task Comments (Faceted by source and phase)",
x = "PC4",
y = "PC3",
)
lm(main_df$human_BE_prop ~ main_df$median_PC1)

85
p2/quest/neurobiber_PCA_analysis.R
View File

@ -1,17 +1,88 @@
library(tidyverse)
library(dplyr)
neurobiber_description_pca_csv <-"~/p2/quest/101325_description_PCA_df.csv"
neurobiber_description_pca_df <- read.csv(neurobiber_description_pca_csv , header = TRUE) |> mutate(comment_text = text)
#neurobiber_description_pca_csv <-"~/p2/quest/101325_description_PCA_df.csv"
#neurobiber_description_pca_df <- read.csv(neurobiber_description_pca_csv , header = TRUE) |> mutate(comment_text = text)
neurobiber_subcomment_pca_csv <-"~/p2/quest/101325_subcomment_PCA_df.csv"
neurobiber_subcomment_pca_df <- read.csv(neurobiber_subcomment_pca_csv , header = TRUE) |> mutate(comment_text = text)
#neurobiber_subcomment_pca_csv <-"~/p2/quest/101325_subcomment_PCA_df.csv"
#neurobiber_subcomment_pca_df <- read.csv(neurobiber_subcomment_pca_csv , header = TRUE) |> mutate(comment_text = text)
pca_csv <- "~/p2/quest/102025_total_pca_df.csv"
pca_df <- read.csv(pca_csv , header = TRUE) |> mutate(comment_text = text)
#pca_csv <- "~/p2/quest/102025_total_pca_df.csv"
#pca_df <- read.csv(pca_csv , header = TRUE) |> mutate(comment_text = text)
main_csv <- "~/analysis_data/100625_unified_w_affil.csv"
main_csv <- "~/analysis_data/102725_unified.csv"
main_df <- read.csv(main_csv , header = TRUE)
main_df <- main_df |>
mutate(
comment_wordcount = as.integer(str_count(replace_na(as.character(comment_text), ""), "\\S+"))
)
description_df <- main_df |>
filter(comment_type == "task_description")
replies_df <- main_df |>
filter(comment_type == "task_subcomment") |>
filter(isGerritBot != TRUE) |>
left_join(
description_df,
by="TaskPHID"
)
ggplot(replies_df, aes(x = autho, y = PC3, fill = comment_type)) +
facet_grid(source~phase, scales="fixed") +
geom_point(shape = 21, alpha=0.3, size=2) +
xlim(-30, 30) +
ylim(-30, 30) +
scale_fill_viridis_d() +
theme_minimal() +
labs(
title = "PCs for Task Comments (Faceted by source and phase)",
x = "PC4",
y = "PC3",
)
replies_df |>
ggplot(aes(
x = as.factor(author_closer.y), # x-axis grouping
y = PC1.x,
fill = reso
)) +
ylim(-30, 30) +
geom_boxplot(alpha = 0.7, position = position_dodge(width = 0.9)) +
facet_grid(. ~ source.x, scales = "fixed") +
scale_fill_viridis_d() +
theme_minimal() +
labs(
title = "Boxplot of PC4",
x = "Comment_type",
y = "PC4",
fill = "isAuthorWMF?"
)
description_df |>
ggplot(aes(
x = as.factor(author_closer), # x-axis grouping
y = PC4,
fill = resolution_outcome
)) +
facet_grid( ~ source, scales = "fixed") +
ylim(-40, 40) +
geom_boxplot(alpha = 0.7, position = position_dodge(width = 0.9)) +
scale_fill_viridis_d() +
theme_minimal() +
labs(
title = "Boxplot of PC4",
x = "Comment_type",
y = "PC4",
fill = "isAuthorWMF?"
)
main_df <- main_df |>
select(TaskPHID, AuthorPHID, date_created, comment_text, isAuthorWMF, isGerritBot, resolution_outcome, task_title, priority)
# Join main_df to neurobiber_description_pca_df