updating commit activity modeling

commit_analysis/case1/0312_ve_core_event_ba_ranefs.csv

@@ -1,9 +0,0 @@
-"","effect","group","level","term","estimate","std.error","conf.low","conf.high"
-"1","ran_vals","commit_type","other_commit_count","(Intercept)",0.364855157809777,0.0448820171494965,0.276888020643255,0.4528222949763
-"2","ran_vals","commit_type","wikimedia_commit_count","(Intercept)",-0.364855157809562,0.0597894025576017,-0.482040233479628,-0.247670082139495
-"3","ran_vals","commit_type","other_commit_count","before_after",0.352924544901635,0.0434143938187066,0.267833896606332,0.438015193196938
-"4","ran_vals","commit_type","wikimedia_commit_count","before_after",-0.352924544901426,0.0578343138752091,-0.466277717167421,-0.239571372635431
-"5","ran_vals","commit_type","other_commit_count","relative_week",0.020534807942067,0.00252605338611267,0.0155838342822607,0.0254857816018733
-"6","ran_vals","commit_type","wikimedia_commit_count","relative_week",-0.0205348079420548,0.00336507207192973,-0.0271302280084187,-0.013939387875691
-"7","ran_vals","commit_type","other_commit_count","before_after:relative_week",-0.0610245395646896,0.00750682666355926,-0.0757376494634507,-0.0463114296659284
-"8","ran_vals","commit_type","wikimedia_commit_count","before_after:relative_week",0.0610245395646535,0.0100001896117208,0.0414245280871091,0.0806245510421979

commit_analysis/case1/0312_ve_core_testing_ba_ranefs.csv

@@ -1,9 +0,0 @@
-"","effect","group","level","term","estimate","std.error","conf.low","conf.high"
-"1","ran_vals","commit_type","other_commit_count","(Intercept)",0,0,0,0
-"2","ran_vals","commit_type","wikimedia_commit_count","(Intercept)",0,0,0,0
-"3","ran_vals","commit_type","other_commit_count","before_after",0.210100230315908,0.0587264628069004,0.0949984782749522,0.325201982356864
-"4","ran_vals","commit_type","wikimedia_commit_count","before_after",-0.210100230315902,0.0620161983828405,-0.331649745604361,-0.0885507150274436
-"5","ran_vals","commit_type","other_commit_count","relative_week",0.000197187070261459,5.51170240557305e-05,8.91596881771995e-05,0.000305214452345719
-"6","ran_vals","commit_type","wikimedia_commit_count","relative_week",-0.000197187070261453,5.82045661455467e-05,-0.000311265923642504,-8.31082168804025e-05
-"7","ran_vals","commit_type","other_commit_count","before_after:relative_week",0.00296979606367595,0.000830106744433257,0.00134281674126297,0.00459677538608893
-"8","ran_vals","commit_type","wikimedia_commit_count","before_after:relative_week",-0.00296979606367587,0.00087660761567573,-0.00468791541897383,-0.00125167670837791

commit_analysis/case1/0312_ve_ve_event_ba_ranefs.csv

@@ -1,9 +0,0 @@
-"","effect","group","level","term","estimate","std.error","conf.low","conf.high"
-"1","ran_vals","commit_type","other_commit_count","(Intercept)",0.19142744594005,0.141824355514081,-0.0865431829981536,0.469398074878253
-"2","ran_vals","commit_type","wikimedia_commit_count","(Intercept)",-0.191427445940039,0.144215326442934,-0.474084291786877,0.091229399906799
-"3","ran_vals","commit_type","other_commit_count","before_after",-0.0755920578722042,0.0560044816831836,-0.185358824944077,0.0341747091996687
-"4","ran_vals","commit_type","wikimedia_commit_count","before_after",0.0755920578721999,0.0569486431221849,-0.0360252316157072,0.187209347360107
-"5","ran_vals","commit_type","other_commit_count","relative_week",0.0275202739990639,0.0203891615662596,-0.0124417483457733,0.0674822963439011
-"6","ran_vals","commit_type","wikimedia_commit_count","relative_week",-0.0275202739990623,0.0207328958451281,-0.0681560031507335,0.0131154551526089
-"7","ran_vals","commit_type","other_commit_count","before_after:relative_week",-0.030560534526985,0.0226416232645461,-0.0749373006770195,0.0138162316230495
-"8","ran_vals","commit_type","wikimedia_commit_count","before_after:relative_week",0.0305605345269833,0.0230233310665026,-0.014564365167504,0.0756854342214706

commit_analysis/case1/0312_ve_ve_testing_ba_ranefs.csv

@@ -1,9 +0,0 @@
-"","effect","group","level","term","estimate","std.error","conf.low","conf.high"
-"1","ran_vals","commit_type","other_commit_count","(Intercept)",0.00187325921107161,0.00396272055679488,-0.00589353036104286,0.00964004878318609
-"2","ran_vals","commit_type","wikimedia_commit_count","(Intercept)",-0.00187325921107161,0.00396516815152807,-0.00964484598071188,0.00589832755856866
-"3","ran_vals","commit_type","other_commit_count","before_after",-0.00139801727712659,0.00295738666067831,-0.00719438862041525,0.00439835406616207
-"4","ran_vals","commit_type","wikimedia_commit_count","before_after",0.00139801727712659,0.00295921330575988,-0.00440193422473448,0.00719796877898766
-"5","ran_vals","commit_type","other_commit_count","relative_week",-8.07243856905614e-05,0.000170765986989873,-0.000415419569975148,0.000253970798594025
-"6","ran_vals","commit_type","wikimedia_commit_count","relative_week",8.07243856905613e-05,0.000170871460969434,-0.00025417752379527,0.000415626295176393
-"7","ran_vals","commit_type","other_commit_count","before_after:relative_week",0.00635509368794709,0.0134436571769782,-0.0199939901994337,0.0327041775753279
-"8","ran_vals","commit_type","wikimedia_commit_count","before_after:relative_week",-0.00635509368794708,0.0134519607247868,-0.0327204522299765,0.0200102648540824

commit_analysis/case1/0314_ve_core_event_ba_ranefs.csv

@@ -0,0 +1,9 @@
+"","effect","group","level","term","estimate","std.error","conf.low","conf.high"
+"1","ran_vals","commit_type","other_commit_count","(Intercept)",0.314596203409928,0.0371497302009551,0.241784070180676,0.38740833663918
+"2","ran_vals","commit_type","wikimedia_commit_count","(Intercept)",-0.314596203409886,0.0616199530813656,-0.435369092178411,-0.193823314641362
+"3","ran_vals","commit_type","other_commit_count","before_after",0.288063179135179,0.034016524269414,0.221392016687895,0.354734341582463
+"4","ran_vals","commit_type","wikimedia_commit_count","before_after",-0.28806317913514,0.0564229300672885,-0.398650089969248,-0.177476268301033
+"5","ran_vals","commit_type","other_commit_count","relative_week",0.0251304693578528,0.00296758241235654,0.0193141147084795,0.0309468240072261
+"6","ran_vals","commit_type","wikimedia_commit_count","relative_week",-0.0251304693578494,0.00492230461850199,-0.0347780091310485,-0.0154829295846504
+"7","ran_vals","commit_type","other_commit_count","before_after:relative_week",-0.00311874402222872,0.000368283244866931,-0.00384056591827745,-0.00239692212617999
+"8","ran_vals","commit_type","wikimedia_commit_count","before_after:relative_week",0.0031187440222283,0.000610868361544776,0.00192146403430555,0.00431602401015106

commit_analysis/case1/0314_ve_core_testing_ba_ranefs.csv

@@ -0,0 +1,9 @@
+"","effect","group","level","term","estimate","std.error","conf.low","conf.high"
+"1","ran_vals","commit_type","other_commit_count","(Intercept)",0.0720020610818892,0.0243405041188196,0.0242955496434539,0.119708572520325
+"2","ran_vals","commit_type","wikimedia_commit_count","(Intercept)",-0.0720020610818955,0.0266280623428783,-0.124192104252024,-0.0198120179117667
+"3","ran_vals","commit_type","other_commit_count","before_after",0.0527000932779371,0.0178154183719228,0.0177825148994553,0.0876176716564189
+"4","ran_vals","commit_type","wikimedia_commit_count","before_after",-0.0527000932779416,0.019489738934506,-0.0908992796576615,-0.0145009068982218
+"5","ran_vals","commit_type","other_commit_count","relative_week",0.016636367754247,0.00562397204296351,0.00561358509997836,0.0276591504085156
+"6","ran_vals","commit_type","wikimedia_commit_count","relative_week",-0.0166363677542484,0.00615252163225638,-0.0286950885675745,-0.00457764694092234
+"7","ran_vals","commit_type","other_commit_count","before_after:relative_week",0.0340353312009155,0.0115057417222755,0.0114844918098356,0.0565861705919954
+"8","ran_vals","commit_type","wikimedia_commit_count","before_after:relative_week",-0.0340353312009184,0.0125870691275504,-0.0587055333618332,-0.00936512904000372

commit_analysis/case1/0314_ve_ve_event_ba_ranefs.csv

@@ -0,0 +1,9 @@
+"","effect","group","level","term","estimate","std.error","conf.low","conf.high"
+"1","ran_vals","commit_type","other_commit_count","(Intercept)",0.406936293894797,0.056676800683778,0.295851805795637,0.518020781993957
+"2","ran_vals","commit_type","wikimedia_commit_count","(Intercept)",-0.406936293894832,0.085077030178694,-0.573684208956699,-0.240188378832964
+"3","ran_vals","commit_type","other_commit_count","before_after",0.00348311531536142,0.000501125210636354,0.00250092795076911,0.00446530267995372
+"4","ran_vals","commit_type","wikimedia_commit_count","before_after",-0.00348311531536171,0.000738964571298917,-0.00493145926095867,-0.00203477136976476
+"5","ran_vals","commit_type","other_commit_count","relative_week",-0.0163312584538643,0.00227482472962926,-0.0207898329950787,-0.0118726839126499
+"6","ran_vals","commit_type","wikimedia_commit_count","relative_week",0.0163312584538657,0.00341450238007504,0.00963895676379232,0.0230235601439391
+"7","ran_vals","commit_type","other_commit_count","before_after:relative_week",-0.0815383698564193,0.0113564215132035,-0.103796547015554,-0.0592801926972846
+"8","ran_vals","commit_type","wikimedia_commit_count","before_after:relative_week",0.0815383698564263,0.017047008558442,0.0481268470377339,0.114949892675119

commit_analysis/case1/0314_ve_ve_testing_ba_ranefs.csv

@@ -0,0 +1,9 @@
+"","effect","group","level","term","estimate","std.error","conf.low","conf.high"
+"1","ran_vals","commit_type","other_commit_count","(Intercept)",0.0448635301708636,0.0600660780690559,-0.072863819537057,0.162590879878784
+"2","ran_vals","commit_type","wikimedia_commit_count","(Intercept)",-0.0448635301708634,0.0603190867086949,-0.163086767700254,0.0733597073585272
+"3","ran_vals","commit_type","other_commit_count","before_after",0.00105436348083387,0.00141164850560976,-0.00171241674899104,0.00382114371065878
+"4","ran_vals","commit_type","wikimedia_commit_count","before_after",-0.00105436348083386,0.00141759460109302,-0.00383279784365461,0.00172407088198689
+"5","ran_vals","commit_type","other_commit_count","relative_week",-0.00804793108391088,0.0107750695316085,-0.0291666792967783,0.0130708171289566
+"6","ran_vals","commit_type","wikimedia_commit_count","relative_week",0.00804793108391084,0.0108204559754881,-0.013159772924347,0.0292556350921687
+"7","ran_vals","commit_type","other_commit_count","before_after:relative_week",0.0379986937948052,0.0508750091414543,-0.0617144918355913,0.137711879425202
+"8","ran_vals","commit_type","wikimedia_commit_count","before_after:relative_week",-0.0379986937948049,0.0510893034199067,-0.138131888493061,0.0621345009034511

commit_analysis/commit_plotting.R

@@ -7,8 +7,8 @@ input_df$nonbot_commit_count <- input_df$commit_count - input_df$bot_commit_coun
 window_num <- 12
 intermediate_df <- input_df |>
   mutate(nonbot_commit_count = commit_count - bot_commit_count)|>
-  mutate(other_commit_count = nonbot_commit_count - mediawiki_dev_commit_count - wikia_commit_count - wikimedia_commit_count) |>
-  mutate(wikimedia_commit_count = wikimedia_commit_count + mediawiki_dev_commit_count + wikia_commit_count) |>
+  mutate(other_commit_count = nonbot_commit_count - mediawiki_dev_commit_count - wikimedia_commit_count) |>
+  mutate(wikimedia_commit_count = wikimedia_commit_count + mediawiki_dev_commit_count) |>
   dplyr::select(-mediawiki_dev_commit_count) |>
   dplyr::select(-wikia_commit_count) |>
   filter(relative_week >= (- window_num) & relative_week <= (window_num)) 

commit_analysis/matched_rdd_models.R

@@ -1,12 +1,12 @@
 library(tidyverse)
 library(dplyr)
-entest_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/en-testing_0312_mediawiki_core_weekly_commit_count_data.csv"
+entest_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/en-testing_0312_extensions_ve_weekly_commit_count_data.csv"
 entest_df <- read.csv(entest_fp, header = TRUE) |> mutate(rd_event = "en-testing")
 
-widetest_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/wide-testing_0312_mediawiki_core_weekly_commit_count_data.csv"
+widetest_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/wide-testing_0312_extensions_ve_weekly_commit_count_data.csv"
 widetest_df <- read.csv(widetest_fp, header = TRUE) |> mutate(rd_event = "wide-testing")
 
-event_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/event_0312_mediawiki_core_weekly_commit_count_data.csv"
+event_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/event_0312_extensions_ve_weekly_commit_count_data.csv"
 event_df <- read.csv(event_fp, header = TRUE) |> mutate(rd_event = "default")
 
 #input_df <- bind_rows(entest_df, widetest_df, event_df)
@@ -16,8 +16,8 @@ input_df <- bind_rows(entest_df, widetest_df)
 
 input_df <- input_df |>
   mutate(nonbot_commit_count = commit_count - bot_commit_count)|>
-  mutate(other_commit_count = nonbot_commit_count - mediawiki_dev_commit_count - wikia_commit_count - wikimedia_commit_count) |>
-  mutate(wikimedia_commit_count = wikimedia_commit_count + mediawiki_dev_commit_count + wikia_commit_count) |>
+  mutate(other_commit_count = nonbot_commit_count - mediawiki_dev_commit_count - wikimedia_commit_count) |>
+  mutate(wikimedia_commit_count = wikimedia_commit_count + mediawiki_dev_commit_count) |>
   dplyr::select(-mediawiki_dev_commit_count) |>
   dplyr::select(-wikia_commit_count)
 
@@ -37,27 +37,36 @@ intermediate_long_df <- long_df |>
   mutate(log_commits = log1p(lengthened_commit_count))|>
   mutate(scaled_long_commits = lengthened_commit_count / 10) 
 
-library(rdd)
+#library(rdd)
+library(rdrobust)
 
 var(intermediate_long_df$lengthened_commit_count) 
 mean(intermediate_long_df$lengthened_commit_count) 
 median(intermediate_long_df$lengthened_commit_count) 
 
-get_optimal_bandwidth <- function(df){
-  bw <- tryCatch({
-    IKbandwidth(df$relative_week, df$lengthened_commit_count, cutpoint = 0, verbose = FALSE, kernel = "triangular")
-  }, error = function(e) {
-    NA
-  })
-}
+#bw <- rdbwselect(intermediate_long_df$lengthened_commit_count, 
+#           intermediate_long_df$relative_week, 
+#           bwselect = "mserd")
+#bw['bws']
+
+#get_optimal_bandwidth <- function(df){
+#  bw <- tryCatch({
+#    IKbandwidth(df$relative_week, df$lengthened_commit_count, cutpoint = 0, verbose = TRUE, kernel = "triangular")
+#  }, error = function(e) {
+#    NA
+#  })
+#}
 
 intermediate_long_df <- intermediate_long_df |>
   filter(commit_type != "unaff_ft_commit_count")
 
-optimal_bandwidth <- get_optimal_bandwidth(intermediate_long_df)
+#optimal_bandwidth <- get_optimal_bandwidth(intermediate_long_df)
+#IKbandwidth(intermediate_long_df$relative_week, 
+#            intermediate_long_df$lengthened_commit_count, 
+#            cutpoint = 0, verbose = TRUE, kernel = "triangular")
 
 library(dplyr)
-window_num <- 12
+window_num <- 4
 final_long_df <- intermediate_long_df |>
   filter(relative_week >= (-window_num) & relative_week <= window_num) |>
   filter(commit_type != "unaff_ft_commit_count")
@@ -76,7 +85,7 @@ mlm <- glmer.nb(lengthened_commit_count ~ before_after*relative_week +
                 data=final_long_df)
 #(before_after*relative_week|rd_event)
 
-saveRDS(mlm, file = "0312_ve_core_event_commits_mlm.rds")
+saveRDS(mlm, file = "0314_ve_ve_testing_commits_mlm.rds")
 #mlm <- readRDS("commit_analysis/case1/0312_core_ve_testing_commits_mlm.rds")
 summary(mlm)
 qqnorm(residuals(mlm))
@@ -88,7 +97,7 @@ library(broom.mixed)
 library(ggplot2)
 condvals <- broom.mixed::tidy(mlm, effects = "ran_vals", conf.int = TRUE)
 glmer_ranef_ba <- condvals
-write.csv(glmer_ranef_ba, "0312_ve_core_event_ba_ranefs.csv")
+write.csv(glmer_ranef_ba, "0314_ve_ve_testing_ba_ranefs.csv")
 
 
 

commit_analysis/testing-share-plotting.R

@@ -1,11 +1,11 @@
 library(tidyverse)
-entest_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/en-testing_0312_extensions_ve_weekly_commit_count_data.csv"
+entest_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/en-testing_0312_mediawiki_core_weekly_commit_count_data.csv"
 entest_df <- read.csv(entest_fp, header = TRUE) |> mutate(rd_event = "en-testing")
 
-widetest_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/wide-testing_0312_extensions_ve_weekly_commit_count_data.csv"
+widetest_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/wide-testing_0312_mediawiki_core_weekly_commit_count_data.csv"
 widetest_df <- read.csv(widetest_fp, header = TRUE) |> mutate(rd_event = "wide-testing")
 
-event_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/event_0312_extensions_ve_weekly_commit_count_data.csv"
+event_fp <-"/mmfs1/gscratch/comdata/users/mjilg/mw-repo-lifecycles/case1/event_0312_mediawiki_core_weekly_commit_count_data.csv"
 event_df <- read.csv(event_fp, header = TRUE) |> mutate(rd_event = "default")
 
 #input_df <- bind_rows(entest_df, widetest_df, event_df)
@@ -15,8 +15,8 @@ input_df <- bind_rows(entest_df, widetest_df)
 
 input_df <- input_df |>
   mutate(nonbot_commit_count = commit_count - bot_commit_count)|>
-  mutate(other_commit_count = nonbot_commit_count - mediawiki_dev_commit_count - wikia_commit_count - wikimedia_commit_count) |>
-  mutate(wikimedia_commit_count = wikimedia_commit_count + mediawiki_dev_commit_count + wikia_commit_count) |>
+  mutate(other_commit_count = nonbot_commit_count - mediawiki_dev_commit_count - wikimedia_commit_count) |>
+  mutate(wikimedia_commit_count = wikimedia_commit_count + mediawiki_dev_commit_count) |>
   dplyr::select(-mediawiki_dev_commit_count) |>
   dplyr::select(-wikia_commit_count)
 
@@ -36,7 +36,7 @@ intermediate_long_df <- long_df |>
   mutate(scaled_long_commits = lengthened_commit_count / 10) 
 
 
-window_num <- 8
+window_num <- 4
 final_long_df <- intermediate_long_df |>
   filter(relative_week >= (- window_num) & relative_week <= (window_num)) 
 
@@ -52,17 +52,17 @@ commit_plot <- final_long_df |>
               linetype = rd_event)) +
   geom_line() +
   geom_point() + 
-  labs(x = "Relative Week", y = "Nonbot Commits", linetype = "Testing Event", color="Commit Author Affiliation") +
+  labs(x = "Relative Week", y = "Commits", linetype = "Testing Event", color="Commit Author Affiliation") +
   scale_color_manual(values = affiliationColors,
                      labels = c("other_commit_count" = "Unaffiliated", "wikimedia_commit_count" = "WMF Affiliated")) + 
-  scale_linetype_discrete(labels = c("enwiki testing (2012-12-12)", "wide testing (2013-04-25)")) + 
-  ggtitle("mw-core New Contributors Commits Around Opt-In Testing Events (by Affiliation)") +
+  scale_linetype_discrete(labels = c("enwiki testing (2012-12-11)", "wide testing (2013-04-25)")) + 
+  ggtitle("mw-core Commits Around Opt-In Testing Events (by Affiliation)") +
   theme_bw() + 
   theme(legend.position = "top")
 commit_plot
 
 
-ggsave(filename = "0312-core-ve-testing-new-commits.png", plot = commit_plot, width = 15, height = 9, dpi = 800)
+ggsave(filename = "0314-ve-core-testing-new-commits.png", plot = commit_plot, width = 15, height = 9, dpi = 800)
 
 total_commit_plot <- final_long_df |>
   ggplot(aes(x=relative_week,