User:Vanished user adhmfdfmykrdyr/Article rank and article date

An issue arose on meta recently, which suggested there was little value in having a historical public record of news stories because people will not access them. The implication was that older English (and other project) Wikinews content would not be accessed for use by others seeking information. Ignoring for the moment that English Wikipedia relies rather heavily on historical news reports both for information and establishing notability, this idea needs a bit of debunking since it is not true.

Article traffic is dependent on a number of factors. For English Wikipedia, this includes the geographic topic (with preference given for North American, Australian, European and Asian content), a general topic focus and other variables. Date of creation alone does not explain traffic patterns.

That said, a quick and dirty analysis involving 50 English Wikinews articles. They included recent English Wikinews featured articles, articles listed as popular now, articles from wacky news, articles from the Real Madrid category, articles from the obituaries category and a few random articles. Included in this sample were 1 article from 2005, 1 article from 2006, 4 articles from 2007, 3 articles from 2008, 4 articles from 2009, 6 articles from 2010, 6 articles from 2011, 9 articles from 2012 and 16 articles from 2013. The English Wikinews rank for these articles was found for the period between July 1 and July 28, 2013.

Bearing in mind that English Wikinews has nearly 20,000 total articles and an even larger number of non-article pages which are counted in the ranks, the ranking for English Wikinews appears to go to around 10,000 most popular articles. In the sample, 4 articles from 2013 were unranked, and one article from 2007 and 2008 were unranked. For the purpose of analysis, these have all been ranked 10,001.

Based on average, the most popular year for articles are in the following order: 2005, 2006, 2012, 2009, 2010, 2013, 2011, 2007 and 2008. This order changes a fair bit when median is used instead: 2013, 2005, 2006, 2012, 2007, 2009, 2010, 2011 and 2008. The difference in ranking for average and median supports the conclusion that article age is not a particularly relevant variable when assessing the relative article popularity on English Wikinews.

Any conclusion that English Wikinews traffic validates a conclusion that a historical archive of open access news sources is not valuable needs to bring their own data to the table. A simple, quick and dirty data anlaysis proves this is not true. Further, a more extensive data analysis would likely show what variables are more likely to be predictors of current traffic patterns to older articles.


 * Raw data

Average and median table

Article table