Github sampling methodology - do event types correlate with overall participation?

missaugustina commented 7 years ago

update: Moved original to Issue #67

Turns out the participation rate isn't necessary for this phase beyond establishing a basis for event type analysis. We can use other parameters to evaluate event types, the accuracy of the participation rate isn't necessary for this.

[x] Split old report into 3 seperate reports
[x] Initial events analysis that led to participation rate idea
[x] Participation Rate report that pulls samples by event type and proves correlation

missaugustina commented 7 years ago

Updated previous events analysis report and split into 2 reports with a separate one for participation rate. First step here is going to be to collect the analysis I did on the initial participation rate stuff and summarize my findings, then move to doing the recalc.

missaugustina commented 7 years ago

Remaining work for the Participation report:

[x] Fix low participation samples
[x] only study list per above results, remove any that aren't from that list
[x] compare population rating samples with population samples
[x] group by participation rate correlation
[x] show participation rate, event type distribution for repos in sample, num_events per repo frequency, num_actor per repo frequency

missaugustina commented 7 years ago

Overall population

[x] analysis
[x] write up/summary

Low participation population

[x] Low Participation analysis
[x] Low Participation write up/summary

Medium participation population

[x] Medium Participation analysis
[x] Medium Participation write up/summary

High participation population

[x] High Participation analysis
[x] High Participation write up/summary

All participation rates event types

[x] Push events (all) - participation rate between samples
[x] Push events (all) - participation rate vs population
[x] Push events (all) - event types between samples
[x] Push events (all) - event types vs population
[x] Push events (all) - events per repo between samples
[x] Push events (all) - events per repo vs population
[x] Push events (all) - actors per repo between samples
[x] Push events (all) - actors per repo vs population
[x] write up/summary

High participation rates event types

[x] Watch events (High) - participation rate between samples
[x] Watch events (High) - actors per repo between samples
[x] Issue Comment events (High) - participation rate between samples
[x] Issue Comment events (High) - actors per repo between samples
[x] Pull Request events (High) - participation rate between samples
[x] Pull Request events (High) - actors per repo between samples
[x] Fork events (High) - participation rate between samples
[x] Fork events (High) - actors per repo between samples
[x] Issues events (High) - participation rate between samples
[x] Issues events (High) - actors per repo between samples
[x] Pull Request Review Comment events (High) - participation rate between samples
[x] Pull Request Review Comment events (High) - actors per repo between samples
[x] Compare actors per repo between event types
[x] write up/summary

Medium participation rates event types

[x] Member events (Medium) - participation rate between samples
[x] Member events (Medium) - actors per repo between samples
[x] Release events (Medium) - participation rate between samples
[x] Release events (Medium) - actors per repo between samples
[x] Compare actors per repo between event types
[x] write up/summary

Low participation rates event types

[x] Create events (Low) - participation rate between samples
[x] Create events (Low) - actors per repo between samples
[x] Public events (Low) - participation rate between samples
[x] Public events (Low) - actors per repo between samples
[x] Delete events (Low) - participation rate between samples
[x] Delete events (Low) - actors per repo between samples
[x] Compare actors per repo between event types
[x] write up/summary

Finalize

[x] Write up conclusions
[x] copy edit
[x] export to PDF

missaugustina commented 7 years ago

Based on current analysis, it looks like for this milestone we do not have to adjust this metric. However for the future milestone on calculating participation, we will want to revisit how this metric is calculated.

missaugustina commented 7 years ago

OMG

passes out

countering-bean-counting / bonnyci_ci-plunder

Github sampling methodology - do event types correlate with overall participation? #56