Research talk:Measuring article importance/Work log/2014-10-23

Thursday, October 23, 2014[edit]

I want to gather a set of datasets:

WikiProject importance classifications
Count of inlinks per page
Page view rate per page

WikiProject importance classifications[edit]

See http://quarry.wmflabs.org/query/794 (Article importance classes for English Wikipedia).

> select importance, COUNT(DISTINCT page_id) from importance_classification group by importance;
+------------+-------------------------+
| importance | COUNT(DISTINCT page_id) |
+------------+-------------------------+
| Top        |                   35639 |
| High       |                  144344 |
| Mid        |                  584461 |
| Low        |                 2359667 |
| Unknown    |                 1794587 |
+------------+-------------------------+
5 rows in set (5.79 sec)

> select LEFT(last_update, 4) AS year, COUNT(DISTINCT page_id) from importance_classification group by year;
+------+-------------------------+
| year | COUNT(DISTINCT page_id) |
+------+-------------------------+
| 2006 |                   22813 |
| 2007 |                  166217 |
| 2008 |                  536320 |
| 2009 |                  786836 |
| 2010 |                  681561 |
| 2011 |                 1070609 |
| 2012 |                  802999 |
| 2013 |                  590660 |
| 2014 |                  587419 |
+------+-------------------------+
9 rows in set (6.80 sec)

Interestingly, it appears that many importance ratings are quite old. It's surprising that, for the most part, importance ratings are updated at a consistent rate. I'd suspect that most importance ratings don't change. --Halfak (WMF) (talk) 14:58, 23 October 2014 (UTC)[reply]

Count of inlinks per article[edit]

See http://quarry.wmflabs.org/query/806 (Inlink counts for all English Wikipedia articles).

Query is still running. I'll come back later to post an update. --Halfak (WMF) (talk) 14:58, 23 October 2014 (UTC)[reply]

Page view rate[edit]

This, I need to think carefully about. I'll ping a few people to see if they have worked through the implications yet. --Halfak (WMF) (talk) 14:58, 23 October 2014 (UTC)[reply]