Analytics/Archive/Pageviews/Aggregation

The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.

This page contains historical information. It may be outdated or unreliable.

Project Animal	{wren}
Phabricator Ticket	https://phabricator.wikimedia.org/T100027
ETA	Q1 FY2016

Value Proposition

For Analysts, Product Managers, Researchers and the Wikimedia Community interested in analytics

Who are dissatisfied with a lack of easy to get & granular pageview data

Our Product is a table you can query via SQL

That provides useful breakdowns (dimensions) while maintaining privacy

Unlike our current pageview data dumps.

Deliverable

A table in a database that is query-able using SQL.

Dimension Name	No. of Possible Values	Examples
project	1000	jpwiki \| frwiki
country	300	Japan \| France
access_method	3	desktop \| mobile_web \| mobile_app
agent_type	2	user \| spider
referer_class	3
day	?

Dimension Name
project
country
access_method
agent_type
referer_class
article
day

Privacy

It is possible the pageview count for a particular bucket will be so small it could be used to identify someone because perhaps only one person could have produced a particular pageview. Therefore when pageview counts are lower than X, the database will return 0.

Implementation