Page MenuHomePhabricator

[REQUEST] updated analysis of Wiki Education participants as a share of new active editors
Closed, ResolvedPublic

Description

What's requested:
An update of the analysis of Wiki Education student editors as a portion of new active editors on English Wikipedia and globally, as reported in the Metrics & Insights Q1 2018-19 presentation (which found that about 19% of new active editors are from Wiki Education's Wikipedia Student Program).

I believe the data comes from this analysis and was updated with some more recent data in June 2019. I'm requesting an update that extends through the rest of 2019 and includes the Spring and Fall 2019 students. Having the exact numbers behind the 'rolling average' graph (where I believe the 19% figure came from) would also be very useful.

Why it's requested:
Wiki Education would like to use this data in grant proposals to external funders, and related communications, to emphasize the impact of our programs and the potential for scaling them up even further. By my estimation from measuring the most recent graph, it looks like Wik Education's portion of new active editors increased to over 20% by late 2018 / early 2019. I expect it would creep a bit higher through the rest of 2019, which were our largest Spring and Fall cohorts to date.

I haven't been able to recreate the analysis myself.

When it's requested:
It would be very useful to have updated data by May 2020 or earlier, both so that we can put out timely communication about recent semesters and so that we can include it in our annual plan. For the purpose of grant proposals, the sooner the better.

Event Timeline

Hi @Ragesoss, acknowledging that the Product Analytics team has received this ticket and will review and prioritize it during our next board review meeting, which is on 2019-02-17.

@Ragesoss one of the analysts on my team, @cchen , has started doing this but she's running into issues getting the full list of participants. Previously, Neil scraped the data from the CSVs you have posted, but we're no longer able to do that due to login authentication. Do you have a full list of participant editor IDs that you can send us?

Please let me know - if not, Connie will begin downloading the CSVs manually.

I will remove the login requirement for the participants CSV downloads. We added that to prevent bots from crawling those too often (since some of them take a while and tie up system resources), but I can figure out a better way to solve that problem. I should have it ready later today so that it can be scraped again.

@cchen @kzimmerman @nshahquinn-wmf I've deployed the change, so the CSVs should be scrapable again now.

Thanks @Ragesoss ! just tested the changes in notebook, it works.

Updated the analysis with Spring/Fall 2019 in this repo.

In English Wikipedia, about 19.8% (3,273 out of 16,449) of new active editors are from Wiki Education's Wikipedia Student Program).

Below is a monthly YoY comparison in active editors from Wiki Edu Program between FY2019/2020 (solid line) and FY 2018/2019 (dot line)

editors.png (1×3 px, 160 KB)

Thank you! Are there any big hurdles to me being able to re-run an analysis like this myself in the future?

@cchen Thank you so much! This is awesome!

One follow-up question: Would this be difficult to re-run for all Wikimedia projects, all languages, with a larger set of usernames? I'm on the board of the Wikipedia & Education User Group, and we've discussed collecting usernames from everyone globally to get a sense of the impact the education program has to all projects' new editors. Collecting these usernames would be a fair amount of work though, so I'd prefer to not ask a lot of people to send them if you wouldn't be able to easily run the number globally.

@Ragesoss, yes. One thing need to change is to pull a list of campaign name to add to "campaign_url" in cell 4. and feel free to reach out to me or our team if you run into any issues.

@LiAnna It's doable as long as we have the full list of participants. Are your team looking for an analysis for global number or analysis by projects or languages?

Mark as resolved.

@LiAnna let me know if you want another analysis across projects, we can open a new ticket for that.