Author: hanteng
Description:
As advised by Erik Zachte, two researchers from the Oxford Internet Institute request the aggregated-to-country-code (and also finer aggregated-to-longitude/latitude-point, if possible) traffic data for all available language versions. The data will be used for improved mapping for Wikimedia Traffic Analysis Report, which shall benefit the public understanding of Wikimedia's multilingual development. The resulted maps will be released in copyleft license.
- Existing tables provided by Erik Zachte's Wikimedia Traffic Analysis Report
Very interesting data presentation to show how different languages are accessed across different regions (based on the geoIP categorization).
http://stats.wikimedia.org/wikimedia/squids/SquidReportPageViewsPerLanguageBreakdown.htm
- Proposed enhancement by our pilot mapping
To show how each language version is accessed across the world on the geographic map, which needs more detailed data, especially for those languages with sizable traffic from the "other" categories in the table shown at the link above.
- Two researchers
Dr. Mark Graham does work on analysing patterns in Wikipedia (e.g. http://www.oii.ox.ac.uk/research/projects/?id=66). Mr. Han-Teng Liao has been using the MaxMind geoIP database for mapping the proportional difference between Baidu Baike and Chinese Wikipedia's external/citation links here: http://people.oii.ox.ac.uk/hanteng/2011/09/04/difference-in-proportional-emphasis-baidu-baike-and-chinese-wikipedia-comparison/.)
- Expected outcome
Published maps released in copy-left license to be stored at Wikicommons. Potential academic articles and blogs on the language phenomenon in multilingual
Wikipedia project.
- Researchers' sensitivity to privacy concerns and capacity in modern cartography.
Both researchers realise the sensitivity of IP data, and in no way want to violate user's expectations of privacy (Han-Teng is especially sensitive to this issue have been involved with Human Rights groups for the Internet industry in DC). This is why we don't want to see any IPs, but would very much like to work with aggregated data to the level of country codes at the very least, or the data aggregated to the level of city or even the longitude and latitude points (Dr Mark Graham is an trained geographer with expertise in mapping both offline and online data, as shown in http://www.floatingsheep.org/ ).
Version: unspecified
Severity: normal