Grants:IdeaLab/Wikipedia Metrics for Institutions
What is the problem you're trying to solve?
Institutional engagement with Wikipedia is increasing worldwide by means of Wikipedians-in-Residence, staff editing, events hosting, image and content donations, bot developing and more. From medical research companies like Cochrane and Consumer Reports to cultural heritage libraries and archives, institutions often choose to engage with Wikipedia because they understand how contributing can be a way to share their valuable authority information and fulfill their institutional missions.
Institutional partnerships are important to the Wikimedia Foundation's goals because they can improve the quality of obscure articles that no one else may be able to write- especially now that most of the 'general' articles have been written, and can convince scholars and researchers of Wikipedia's reliability. Institutions also hold millions of photos, documents, and data and are increasingly looking for open platforms to disseminate their holdings online.
Institutions love numbers, and metrics telling institutions what the current coverage of their institutional resources is on Wikipedia and the viewership of these pages would communicate the huge impact that editing would have and increase institutional buy-in. But at the moment institutions choose to partner with Wikipedia for these verbal arguments rather than any statistical data or metrics.
As an outreach-based Wikipedian-in-Residence, I have noticed many institutions asking about Wikipedia metrics and using Google Analytics when available to trace Wikipedia traffic to their sites and the page views of certain pages. Projects which have used metrics, such as William Blueher's Wikipedia editing project at the Metropolitan Museum's Watson Library have also received a lot of attention from their institutions.
Increasing the quality of Wikipedia's metrics for institutions and creating easily accessible pages and introductions to these metrics, I believe would increase institutional partnerships and content donations greatly.
What is your solution?
There will be two main recipients of the grant- a developer and project documentation/outreach affiliate.
1. Develop tools for more easily tagging articles as related to institutions in the way that WikiProjects currently do so that metrics can be gathered on the number of articles/images related to them on WP, article quality metrics and total page views. New ways to tag articles so that page view statistics/article traffic for a group of articles could be calculated quickly will also be developed.
2. Aggregate metrics about institutional coverage on Wikipedia using current tools available and/or developed and other external tools like Google Analytics. These metrics will be compiled and documented for easy viewing on WP. The project will also work on pages for Wikipedia metrics for institutions,aggregating links to metrics tools which work on WP already, giving instructions for documentation and working with Wikipedians in the institutional partnership community to disseminate any new tools created. This effort would add to: Wikipedia:GLAM/Statistics
There are at least two different detail levels; one is to list the aggregate page views within a time period, and another is to give detailed statistics. The first one is most interesting when the effect of some project shall be measured, while the last is most important in a planning phase where you need to know which articles you should focus on. Then there are at least two different ways to specify which pages to include; one is to give a category, and one is to give a list of articles. The category solution is probably much faster when you want to drill down to find interesting articles for further work.
I would propose a special page that can utilize the traffic dumps, either directly or stored in a database table, with some precomputed data (24 hours/1 day, 7 days/1 week, 21days/3 weeks, and 13 weeks/3 months). The precomputed data should be updated each day, there are methods to this fairly efficient. The special page itself should take a category and allow further navigation in the category tree. Each item should have a number of views and a scale to show the page views. The scale would be local for the category, and a reference should be shown on the page. Note also that all redirects should be counted as going to the same article.
A simple generalization would be to allow such a special page to be used for files too, and that should not pose any real problems. In that case the traffic stats should also include places where the file is embedded on a page. There is a WMF-labs tool that does something similar, but only for images. — Jeblad 14:56, 2 November 2014 (UTC)
Welcome, brainstormers! Your feedback on this idea is welcome. Please click the "discussion" link at the top of the page to start the conversation and share your thoughts.