WordPress Agents Exploratory Analysis- Part I- Univariate description - Refactored
Viviane Schneider
08/01/2019
Type of Analysis: Descriptive analysis - Univariate description
Intention of Analysis: 1- Understand how Core WordPress Community coproduce code. 2- Identify potential indicators for coherence analysis.
General Question: Wich are the atributes of coproduction (colunms of dataframe or variables)? Specific Questions: Which type of developers groups exist in WC? It is possible to make some indicator from this data?
Source: Data come from WordPress Report Trac System. URL Source Dataframe: GitHub Repository
Date collection: 04/07/2019.
1.GENERAL ANALYSIS
#READ DATA
library(readr)
library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionlibrary(plotly)## Loading required package: ggplot2##
## Attaching package: 'plotly'## The following object is masked from 'package:ggplot2':
##
## last_plot## The following object is masked from 'package:stats':
##
## filter## The following object is masked from 'package:graphics':
##
## layoutlibrary(tidyr)
TicketW <- read_csv('~/PhD Analysis/1. PhD escriptive exploratory analysis/TicketW.csv')## Parsed with column specification:
## cols(
## id = col_double(),
## Summary = col_character(),
## Status = col_character(),
## Version = col_logical(),
## Owner = col_character(),
## Type = col_character(),
## Priority = col_character(),
## Milestone = col_character(),
## Component = col_character(),
## Severity = col_character(),
## Resolution = col_character(),
## Created = col_character(),
## Modified = col_character(),
## Focuses = col_character(),
## Reporter = col_character(),
## Keywords = col_character()
## )View(TicketW)
dim(TicketW) #dimension## [1] 2333 16TicketW[1:5,] #5 fist lines## # A tibble: 5 x 16
## id Summary Status Version Owner Type Priority Milestone Component
## <dbl> <chr> <chr> <lgl> <chr> <chr> <chr> <chr> <chr>
## 1 24579 Add Dr~ new NA <NA> enha~ high Future R~ Upgrade/~
## 2 30361 Correc~ assig~ NA pento task~ high <NA> General
## 3 32502 Cannot~ new NA <NA> defe~ high <NA> Administ~
## 4 36441 Custom~ new NA <NA> defe~ high Future R~ Customize
## 5 40439 Save p~ assig~ NA mike~ enha~ high 5.3 Media
## # ... with 7 more variables: Severity <chr>, Resolution <chr>,
## # Created <chr>, Modified <chr>, Focuses <chr>, Reporter <chr>,
## # Keywords <chr>summary(TicketW)## id Summary Status Version
## Min. : 5235 Length:2333 Length:2333 Mode:logical
## 1st Qu.:34555 Class :character Class :character NA's:2333
## Median :40511 Mode :character Mode :character
## Mean :38029
## 3rd Qu.:44485
## Max. :47640
## Owner Type Priority
## Length:2333 Length:2333 Length:2333
## Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
##
## Milestone Component Severity
## Length:2333 Length:2333 Length:2333
## Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
##
## Resolution Created Modified
## Length:2333 Length:2333 Length:2333
## Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
##
## Focuses Reporter Keywords
## Length:2333 Length:2333 Length:2333
## Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
## glimpse(TicketW)## Observations: 2,333
## Variables: 16
## $ id <dbl> 24579, 30361, 32502, 36441, 40439, 41292, 41886, 41...
## $ Summary <chr> "Add Drag'n'Drop UI to plugin and theme manual uplo...
## $ Status <chr> "new", "assigned", "new", "new", "assigned", "reope...
## $ Version <lgl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,...
## $ Owner <chr> NA, "pento", NA, NA, "mikeschroder", "jnylen0", "me...
## $ Type <chr> "enhancement", "task (blessed)", "defect (bug)", "d...
## $ Priority <chr> "high", "high", "high", "high", "high", "high", "hi...
## $ Milestone <chr> "Future Release", NA, NA, "Future Release", "5.3", ...
## $ Component <chr> "Upgrade/Install", "General", "Administration", "Cu...
## $ Severity <chr> "normal", "normal", "major", "normal", "normal", "n...
## $ Resolution <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,...
## $ Created <chr> "06/14/2013 05:03:38 PM", "11/17/2014 12:10:55 PM",...
## $ Modified <chr> "04/12/2019 11:04:54 AM", "06/04/2019 07:42:28 PM",...
## $ Focuses <chr> NA, "ui, administration", NA, NA, "ui", NA, NA, NA,...
## $ Reporter <chr> "tw2113", "pento", "ryan", "azaozz", "mikeschroder"...
## $ Keywords <chr> "ui-feedback ux-feedback needs-patch shiny-updates"...Function unianalysis() - Transform vector into a data frame with frequency of levels and proportion
# Transform vector into a data frame with frequency of levels and proportion
.Unianalysis = function (x) {
y <- as.data.frame(as.table(table(x)))
y <- mutate(y, proportion = prop.table(y$Freq) *100)#Proportion
y <- arrange(y, desc(y$Freq))
return(y)
}1.1 Variables related with members Analysis Goal: Find which variables have hight variability, and find a line of cut, in order to use into Bivariate Analysis.
# Var1
Status<-.Unianalysis(TicketW$Status)
#Var2
TType<-.Unianalysis(TicketW$Type)
#Var3
Priority<-.Unianalysis(TicketW$Priority)
#Var4
Milestone<-.Unianalysis(TicketW$Milestone)
#Var5
Component<-.Unianalysis(TicketW$Component)
Component = filter(Component, Freq>90) # Filter components with more than 100 tickets
sum(Component$Freq) #Total Tickets into most frequent components (more than 100 tickets)## [1] 738#Var6
Severity<-.Unianalysis(TicketW$Severity)
#Var7
Focuses<-.Unianalysis(TicketW$Focuses)
Focuses = filter(Focuses, Freq>10) # Filter Focuses with more than 10 tickets
Focuses<-Focuses[order(Focuses$Freq, decreasing = TRUE),]
sum(Focuses$Freq) #Total Tickets into most frequent components (more than 100 tickets)## [1] 832#Var8
Keywords<-.Unianalysis(TicketW$Keywords)
Keywords = filter(Keywords, Freq>8)
Keywords<-Keywords[order(Keywords$Freq, decreasing = TRUE),]
sum(Keywords$Freq)## [1] 1082Functions for establish groups of active (GroupActive()), median (GroupMedian()), or less active (GroupAlien()) members:
#1. Filter Group of agents
.Grouping = function(x, less, more) {
y = filter(x, Freq < less & Freq > more)
x <- y[order(y$Freq, decreasing = TRUE),]
return(x)
}1.2 Variable Members Analysis - Reporters: Reporters are WordPress Community members who find and report a problem from WP Platform, into a Ticket.
Goal: Find type of report member groups (actives, medians, and aliens), and find a line of cut, in order to use into Bivariate Analysis.
#Find Groups and quantity of people for each group
Reporter<-.Unianalysis(TicketW$Reporter)
totalReporter<-nrow(Reporter)
ActiveReporters<-.Grouping(Reporter, 1000, 10)
totalAR<-nrow(ActiveReporters)
MedianReporters<-.Grouping(Reporter, 10, 4)
totalMR<-nrow(MedianReporters)
LessReporters<-.Grouping(Reporter, 4, 0)
totalLR<-nrow(LessReporters)1.3 Variable Member Analysis - Owners: Owners are WordPress Community members who pick up a ticket from WP Platform (sended by a reporter) in order to solve it.
Goal: Find type of owner member groups (actives, medians, and aliens), and find a line of cut, in order to use into Bivariate Analysis.
Owner <-.Unianalysis(TicketW$Owner)
totalOwner <-nrow(Owner)
#Find Groups and quantity of people for each group
ActiveOwner <-.Grouping(Owner, 1000, 10)
totalAO <-nrow(ActiveOwner)
MedianOwner <-.Grouping(Owner, 10, 4)
totalMO <-nrow(MedianOwner)
LessOwner <-.Grouping(Owner, 4, 0)
totalLO <-nrow(LessOwner)1.4 Sum of groups by Reporters and Owners:
#Total members per Reporter
MembersTotal<-c(totalAR,totalMR,totalLR)
CoreGroup<-rbind("Active Reporters","Median Reporters","Alien Reporters")
WPCGroupR<-data.frame(CoreGroup,MembersTotal)
#Total members per Owner
MembersTotal<-c(totalAO,totalMO,totalLO)
CoreGroup<-rbind("Active Owners","Median Owners","Alien Owners")
WPCGroupO<-data.frame(CoreGroup,MembersTotal)
#Total members per every groups
MembersTotal<- c(totalOwner,totalReporter)
CoreGroup <- rbind("Owners","Reporters")
WPCGroups <- data.frame(CoreGroup,MembersTotal)2. DESCRIPTION ANALYSIS REPORT :
2.1 Variables related with members Report: Variables selected to bivariate analysis are Component(+100 tickets per level), Focuses (+1 Ticket per level), Keywords(+9 tickets per level), Type, Status.
library(plotrix)
library(plotly)
library(ggplot2)
library(wordcloud)## Loading required package: RColorBrewer#Graphic Function
#Fun Plot
.Plot_FunPlot = function(x,y) {
fan.plot(x$Freq,
max.span=pi,
labels=paste(x$x, x$Freq, sep=": "),
main=y,ticks=360)
}
#Word Plot
.Plot_word = function(x, num1, num2) {
wordcloud(words = x$x, freq = x$Freq, min.freq = num1,
max.words=num2, random.order=FALSE, rot.per=0.35,
colors=brewer.pal(8, "Dark2"))
}
Status#Var1## x Freq proportion
## 1 new 1844 79.039863
## 2 assigned 215 9.215602
## 3 reopened 116 4.972139
## 4 reviewing 105 4.500643
## 5 accepted 53 2.271753#Graphics
.Plot_FunPlot(Status, "Tickets per Status")
TType# Var2## x Freq proportion
## 1 enhancement 1109 47.535362
## 2 defect (bug) 1012 43.377625
## 3 feature request 168 7.201029
## 4 task (blessed) 44 1.885984#Graphics
.Plot_FunPlot(TType, "Tickets per Type")
Priority# Var3## x Freq proportion
## 1 normal 2265 97.0852979
## 2 low 49 2.1003000
## 3 high 12 0.5143592
## 4 lowest 7 0.3000429#Graphics
.Plot_FunPlot(Priority, "Tickets per Priority")
Milestone#Var4## x Freq proportion
## 1 Awaiting Review 1104 60.19629226
## 2 Future Release 544 29.66194111
## 3 5.3 172 9.37840785
## 4 WordPress.org 13 0.70883315
## 5 5.2.3 1 0.05452563#Graphics
.Plot_FunPlot(Milestone, "Tickets per Milestone")
Component#Var5## x Freq proportion
## 1 General 205 8.786970
## 2 Media 166 7.115302
## 3 Administration 162 6.943849
## 4 Posts, Post Types 114 4.886412
## 5 Users 91 3.900557#Graphic
.Plot_FunPlot(Component, "Tickets per Component (more than 90 tickets)")
Severity #Var6## x Freq proportion
## 1 normal 2186 93.69909987
## 2 minor 91 3.90055722
## 3 major 27 1.15730819
## 4 trivial 19 0.81440206
## 5 critical 9 0.38576940
## 6 blocker 1 0.04286327#Graphics
.Plot_FunPlot(Severity, "Tickets per Severity")
Focuses#Var7## x Freq proportion
## 1 administration 192 19.512195
## 2 ui 103 10.467480
## 3 ui, administration 101 10.264228
## 4 multisite 95 9.654472
## 5 docs 58 5.894309
## 6 performance 55 5.589431
## 7 ui, accessibility 51 5.182927
## 8 javascript 49 4.979675
## 9 template 48 4.878049
## 10 rest-api 30 3.048780
## 11 coding-standards 27 2.743902
## 12 accessibility 23 2.337398#Graphics
pie(Focuses$Freq, main="Focuses", label = paste(Focuses$x, sep=": ", Focuses$Freq), col = rainbow(7))
.Plot_word(Focuses, 20, 200)
Keywords#Var8## x Freq proportion
## 1 needs-patch 266 15.2873563
## 2 has-patch 246 14.1379310
## 3 reporter-feedback 57 3.2758621
## 4 2nd-opinion 55 3.1609195
## 5 has-patch needs-refresh 47 2.7011494
## 6 needs-patch needs-unit-tests 46 2.6436782
## 7 has-patch needs-testing 44 2.5287356
## 8 has-patch dev-feedback 40 2.2988506
## 9 has-patch has-unit-tests 31 1.7816092
## 10 has-patch 2nd-opinion 29 1.6666667
## 11 dev-feedback 28 1.6091954
## 12 close 27 1.5517241
## 13 has-patch needs-unit-tests 24 1.3793103
## 14 needs-patch 2nd-opinion 19 1.0919540
## 15 good-first-bug has-patch 18 1.0344828
## 16 has-patch reporter-feedback 17 0.9770115
## 17 has-screenshots 15 0.8620690
## 18 has-screenshots has-patch 15 0.8620690
## 19 has-patch has-screenshots 14 0.8045977
## 20 dev-feedback needs-patch 13 0.7471264
## 21 needs-patch dev-feedback 13 0.7471264
## 22 2nd-opinion needs-patch 9 0.5172414
## 23 dev-feedback has-patch 9 0.5172414#Graphics
pie(Keywords$Freq, main="Focuses", label = paste(Keywords$x, sep=": ", Keywords$Freq), col = rainbow(7))
.Plot_word(Keywords, 4, 300)## Warning in wordcloud(words = x$x, freq = x$Freq, min.freq = num1, max.words
## = num2, : has-patch could not be fit on page. It will not be plotted.
2.2 Reporter Members Analysis Report The groups of Reporters:
ActiveReporters #Ranking the most active reporters## x Freq proportion
## 1 johnbillion 72 3.0861552
## 2 afercia 53 2.2717531
## 3 johnjamesjacoby 47 2.0145735
## 4 sebastian.pisula 43 1.8431204
## 5 flixos90 42 1.8002572
## 6 danielbachhuber 39 1.6716674
## 7 karmatosed 33 1.4144878
## 8 rmccue 27 1.1573082
## 9 subrataemfluence 25 1.0715817
## 10 westonruter 24 1.0287184
## 11 nacin 23 0.9858551
## 12 desrosj 22 0.9429919
## 13 garrett-eclipse 22 0.9429919
## 14 pbiron 22 0.9429919
## 15 azaozz 21 0.9001286
## 16 dd32 21 0.9001286
## 17 boonebgorges 19 0.8144021
## 18 pento 17 0.7286755
## 19 ramiy 17 0.7286755
## 20 scribu 17 0.7286755
## 21 anevins 16 0.6858123
## 22 dshanske 15 0.6429490
## 23 helen 15 0.6429490
## 24 iseulde 15 0.6429490
## 25 markjaquith 15 0.6429490
## 26 melchoyce 15 0.6429490
## 27 SergeyBiryukov 15 0.6429490
## 28 swissspidy 15 0.6429490
## 29 tazotodua 15 0.6429490
## 30 Presskopp 14 0.6000857
## 31 henry.wright 13 0.5572225
## 32 mark-k 12 0.5143592
## 33 mukesh27 12 0.5143592
## 34 peterwilsoncc 11 0.4714959MedianReporters #Ranking the Median active reporters## x Freq proportion
## 1 ericlewis 8 0.3429061
## 2 iandunn 8 0.3429061
## 3 jeremyfelt 8 0.3429061
## 4 joostdevalk 8 0.3429061
## 5 kjellr 8 0.3429061
## 6 netweb 8 0.3429061
## 7 pavelevap 8 0.3429061
## 8 programmin 8 0.3429061
## 9 alexvorn2 7 0.3000429
## 10 andy 7 0.3000429
## 11 eclare 7 0.3000429
## 12 Frank Klein 7 0.3000429
## 13 jdgrimes 7 0.3000429
## 14 pbearne 7 0.3000429
## 15 soulseekah 7 0.3000429
## 16 wonderboymusic 7 0.3000429
## 17 xkon 7 0.3000429
## 18 dlh 6 0.2571796
## 19 Ipstenu 6 0.2571796
## 20 jonoaldersonwp 6 0.2571796
## 21 jorbin 6 0.2571796
## 22 kraftbj 6 0.2571796
## 23 mor10 6 0.2571796
## 24 smerriman 6 0.2571796
## 25 WraithKenny 6 0.2571796
## 26 yoavf 6 0.2571796
## 27 allancole 5 0.2143163
## 28 BjornW 5 0.2143163
## 29 ComputerGuru 5 0.2143163
## 30 imath 5 0.2143163
## 31 jipmoors 5 0.2143163
## 32 joemcgill 5 0.2143163
## 33 mikejolley 5 0.2143163
## 34 milana_cap 5 0.2143163
## 35 Rarst 5 0.2143163
## 36 rnaby 5 0.2143163LessReporters[c(1:30),c(1:2)] #Ranking the 30 less active Reporters## x Freq
## 1 allendav 3
## 2 arena 3
## 3 chetan200891 3
## 4 chinteshprajapati 3
## 5 danieltj 3
## 6 Denis-de-Bernardy 3
## 7 diddledan 3
## 8 dimadin 3
## 9 dotancohen 3
## 10 duck_ 3
## 11 fliespl 3
## 12 GaryJ 3
## 13 hlashbrooke 3
## 14 ishitaka 3
## 15 jason_the_adams 3
## 16 javorszky 3
## 17 joehoyle 3
## 18 joyously 3
## 19 jtsternberg 3
## 20 keraweb 3
## 21 ketanumretiya030 3
## 22 knutsp 3
## 23 krogsgard 3
## 24 mehulwpos 3
## 25 MikeHansenMe 3
## 26 Mista-Flo 3
## 27 mnelson4 3
## 28 monikarao 3
## 29 mrwweb 3
## 30 mt8.biz 3#Grafic of Top 5 active Reporters
.Plot_FunPlot(ActiveReporters[1:5, 1:2], "5 Top Active Reporters")
.Plot_FunPlot(MedianReporters[1:5, 1:2], "5 Top Median Reporters")
.Plot_FunPlot(LessReporters[1:5, 1:2], "5 Top Less Active Reporters")
2.3 Owner Members Analysis Report: The groups of Owners:
ActiveOwner#Ranking the most active Owners## x Freq proportion
## 1 SergeyBiryukov 67 18.662953
## 2 audrasjb 18 5.013928
## 3 johnbillion 16 4.456825
## 4 chriscct7 15 4.178273
## 5 joemcgill 14 3.899721
## 6 adamsilverstein 13 3.621170MedianOwner#Ranking the Median Owners## x Freq proportion
## 1 desrosj 9 2.506964
## 2 nacin 8 2.228412
## 3 afercia 7 1.949861
## 4 boonebgorges 7 1.949861
## 5 pento 7 1.949861
## 6 azaozz 6 1.671309
## 7 flixos90 6 1.671309
## 8 jeremyfelt 6 1.671309
## 9 rmccue 6 1.671309
## 10 westi 6 1.671309
## 11 westonruter 5 1.392758
## 12 wonderboymusic 5 1.392758LessOwner[c(1:30),c(1:2)]#Ranking the 30 less active Owners## x Freq
## 1 dd32 3
## 2 joehoyle 3
## 3 tz-media 3
## 4 xkon 3
## 5 davidakennedy 2
## 6 dswebsme 2
## 7 ericlewis 2
## 8 garrett-eclipse 2
## 9 ianbelanger 2
## 10 iseulde 2
## 11 jnylen0 2
## 12 joostdevalk 2
## 13 kapeels 2
## 14 markjaquith 2
## 15 schlessera 2
## 16 sorich87 2
## 17 technosailor 2
## 18 valendesigns 2
## 19 aaroncampbell 1
## 20 adam3128 1
## 21 allisonplus 1
## 22 antpb 1
## 23 bassgang 1
## 24 bhargavbhandari90 1
## 25 birgire 1
## 26 danielbachhuber 1
## 27 danielpataki 1
## 28 davidjlaietta 1
## 29 DrewAPicture 1
## 30 duck_ 1#Grafic of Owners
#Grafic of Top 5 active Owners
.Plot_FunPlot(ActiveOwner[1:5, 1:2], "5 Top Active Owners")
.Plot_FunPlot(MedianOwner[1:5, 1:2], "5 Top Median Owners")
.Plot_FunPlot(LessOwner[1:5, 1:2], "5 Top Less Active Owners")
2.4 Comparation of groups by Reporters and Owners:
WPCGroups## CoreGroup MembersTotal
## 1 Owners 116
## 2 Reporters 1028par(mfrow=c(1,2))
barplot(WPCGroupR$MembersTotal,
names.arg=WPCGroupR$MembersTotal,
xlab="Reporter Groups",
ylab="Total Members",
legend=WPCGroupR$CoreGroup,
col=rainbow(8),
main="Reporters per Groups",border="red")
barplot(WPCGroupO$MembersTotal,
names.arg=WPCGroupO$MembersTotal,
xlab="Owners Groups",
ylab="Total Members",
legend=WPCGroupO$CoreGroup,
col=rainbow(8),
main="Owners per Groups",border="red")
# Load ggplot2
library(ggplot2)
library(dplyr)
# library
library(treemap)
# treemap
.Tree_Map = function(x){
group <- paste(x$CoreGroup, x$MembersTotal)
value <- x$MembersTotal
data <- data.frame(group,value)
treemap(data,
index="group",
vSize="value",
type="index")
}
par(mfrow=c(1,3))
.Tree_Map(WPCGroups) #Total Owner and Reporters
.Tree_Map(WPCGroupO) #Groups of Owners 
.Tree_Map(WPCGroupR) #Groups of Reporters
3. POTENTIAL INDICATORS:
- The “Sense of Unity” Dimension indicator:
- Mode_Agents_Rules: In a scale from O to 1, this indicator represents the principals rules the WP members play in the community: active, median and alien. This unformal rules are observed into formal groups: owners and reporters. The Hypothesis is that the Reporters Group and the Owners Group have a similar distribution of tickets per its the respectives sub groups(active, median, and alien).
Hypothesis: \(PropMo = PropMR\)
Means: percentual of tickets into any subgroups into Owner Groups are similar thand percentual of tickets into any subgroups into Reporter Groups
Where PropMo is the proportion of tickets per active, median, and alien groups of Owners, and PropMR is the proportion of tickets per active, median, and alien groups of Reporters.
Hypothesis test:
PropMR<-prop.table(WPCGroupR$MembersTotal)# Proportion
PropMO<-prop.table(WPCGroupO$MembersTotal)# Proportion
similarityOR<-ifelse(PropMO>PropMR,1-(PropMO-PropMR),1-(PropMR-PropMO))
similarityOR<-mean(similarityOR)
similarityOR## [1] 0.9396642Thus, Mode_Agents_Rules indicator is = similarityOR.
4. FINAL CONCLUSIONS:
There are more Reporters than Owners. Maybe because to report a problem in the WorPress Platform is easear than fix a problem. There are a two members, who can been considers the super active menbers of WP Community: SergeyBiryukov and johnbillion. SergeyBiryukov (number 1 of owners) was reponsable to 67 tickets, and also he report 15 Tickets (23o position of reporters). johnbillion (number 1 of owners) reports 72 tickets, and also was responsable for 16 tickets (3o position of owners).
The hypothesis test proof (result: 0.9772515) indicates that there are a strong coesion into formals groups of core: owners and reporters.
Despite the most of tickets have no owner, there are not so many urgent issues. Around 93% of tickets are classified as Severity “normal”, and 97% are classified as Priority “Normal”.
According with this description, are recommended analyse in a bivariable description the folow variables:
Variables for setup a referential of the community (The “Referentiality” Dimension): Documentation of values, general view and beiefs of community
Variables for analyse the “Prescritivity” dimension: Analyse the community referential with “Focuses” Variable, in order to verify if the members produced what they stablished into referential.
Variables for analyse the “Sense of Unity” dimension: “Keywords” Variable, in order to verify the correlation with groups
Variables for analyse the the “Receptivity” dimension: Groups of members, correlation between groups and Status, Type, Milestone variables