Hi everybody
some of you know that I look from time to time into the statistics for
African language wikipedias. Today I made a comparison which makes me
doubt the usefulness of our beautiful wikimedia statistics, at least for
small wikipedias.
I found large differences for the same months in different tables they
publish. Maybe it has to do with different data for different tables,
but then it is too well hidden for an average reader to notice (and thus
not helpful).
I wondered about large differences between May and November 2017 when
looking at
1. https://stats.wikimedia.org/EN_Africa/Sitemap.htm Wednesday May 31,
2017 ( TABLE 1)
2. https://stats.wikimedia.org/EN/Sitemap.htm Thursday November 30,
2017 (TABLE 2)
I compared to 3.
https://stats.wikimedia.org/EN/TablesPageViewsMonthlyCombined.htm (TABLE 3)
*RESULT: for views by hour*(default in tables 1+2); in table 3 I divide
the given monthly rounded figure by 30, then by 24
*Language *
*Nov 2017 Table1*
*/Nov 2017 Table 3/***
*May 2017 Table 2*
*/May 2017 Table 3/***
arz (Egyptian Arabic)
1204
/2668/
4028
/3194/
sw (Swahili)
570
/2361/
1905
/1667/
yo (Yoruba)
468
/381/
1274
/324/
af (Afrikaans)
2560
/2361/
5640
/3056/
This looks to me highly contradictory and at least for smaller
wikipedias the figures are thus useless.
Which is a pity! I try to ask the statistics team.
Kipala-Ingo