Wikimedia Traffic Analysis Report - Wikipedia Page Views Per Country - Breakdown

Monthly requests or daily averages, for period: 1 Jul 2009 - 31 Dec 2013
000 ⇒ k

 This analysis is based on a 1:1000 sampled server log (squids)

 Page Views Per Country - Overview / Breakdown / Trends,    Page Views Per Wikipedia Language - Breakdown

 Page Edits Per Country - Overview / Breakdown / Trends,    Page Edits Per Wikipedia Language - Breakdown

Showing even small percentages (> 0.1%) (read more). Switch to concise version

 Notes on reliability of these data

 Unresolved Bugzilla bugs: 55443

Recently resolved bugs: 46205 (Aug 2013)46289 (Nov 2013)

Portal is

Show index

Countries by population - English Wikipedia

Internet penetration (% of population) - English Wikipedia

Global North South - English Wikipedia

Share is the percentage of requesting ip addresses (out of the global total) which originated from this country
 Further percentages show per country share of page views per Wikipedia visited

Countries are only included if the number of requests in the period exceeds 10000(10 matching records in 1:1000 sampled log)

Wikipedia's are only listed for some country if the share of visitors for that particular country exceeds 0.1%.

The second column displays the actual numbers of records found in the 1:1000 sampled log on which the percentage is based.
Multiply by 1000 for actual page views over the whole period of 0 months.
If the number of records in the sampled log does not reach the (arbitrary) number of 5 per sampled month, the percentage is flagged dark red to extra emphasize high inaccuracy.

Page requests by bots are not included. Also all ip addresses that occur more than once on a given day are discarded for that day.
A few false negatives are taken for granted. Country meta data collected from English Wikipedia. Portal =

Errata: WMF traffic logging service suffered from server capacity problems in Aug/Sep/Oct 2011.
Absolute traffic counts for October 2011 are approximatly 7% too low.
Data loss only occurred during peak hours. It therefore may have had somewhat different impact for traffic from different parts of the world.
and may have also skewed relative figures like share of traffic per browser or operating system.

From mid September till late November squid log records for mobile traffic were in invalid format.
Data could be repaired for logs from mid October onwards. Older logs were no longer available.

In a an unrelated server outage precisely half of traffic to WMF mobile sites was not counted from Oct 16 - Nov 29 (one of two load-balanced servers did not report traffic).
WMF has since improved server monitoring, so that similar outages should be detected and fixed much faster from now on.

Generated on Tue, Dec 24, 2013 15:19
Author:Erik Zachte (
Web site)
Mail: ezachte@### (no spam: ### =
All data on this page are in the public domain.

Note: page may load slower on Microsoft Internet explorer than on other major browsers