When we first saw Mark Graham’s static maps of geo-located Wikipedia articles on the Guardian Data Blog we were fascinated and wanted to see if it was possible to make an interactive version. We were also intrigued by the idea of creating a temporal version, giving a glimpse of Wikipedia’s evolution over the last decade, and mapping the collaborative effort of thousands of contributors.
You can see the project here:
Mapping Wikipedia Timeline →
By introducing the timeline it’s possible to see how the geographic coverage of Wikipedia has developed for each language. The time slider can be animated to show either a cumulative period or a window of fixed duration. The articles appear when the article stub is created.
Mapping Wikipedia Timeline on Vimeo →
Language Density Maps
The following videos show how the density of articles for English, French and Arabic have accumulated over time.
Not suprisingly the English language version of Wikipedia is the largest and most developed with over 700,000 geolocated articles. This is the heaviest for Mapping Wikipedia to load, taking some minutes for the full dataset. If you don’t have time to view the interactive it’s well worth looking at this and the other videos on our Vimeo channel.
The location of articles emanates from the English speaking parts of the world and spreads over time. In some regions it appears that the density mirrors population, it is possible to make out the larger desert areas of the earth, but many other places with huge populations are barely represented whilst parts of Antarctica, for instance, have vast numbers of articles. And this is where it gets a little confusing, as not all articles are created by individual Wikipedia contributors in person. It is possible to see which article stubs are likely to be created by bots, automatically injecting basic data for locations from public domain sources, as is the case with the Antarctica data. Bot generated stubs tend to appear all at once in large regional blocks. If you switch the Map to Word Count there’s a clue as to which articles have remained as stubs.
The French version of wikipedia was created on or around 11 May 2001 with the Arabic version coming some time after.
At the time the project data was collated the French Wikipedia had a little over 220,000 articles while the Arabic version had around 24,000 articles. When viewed together it is clear how uneven the distribution of articles are across the different Wikipedia languages.
Language Word Count Maps
The following 3 videos show English, French and Arabic again, this time with Word Count selected and a date range of 4 months. Articles appear when the stub is created, and disappear again when the time slider no longer covers the creation date. This view allows us to see the patterning of article production at specific times. (Small changes are not obscured by older articles as in the cumulative animations above).
Mapping Wikipedia – English – Word Count – 4 month range
Mapping Wikipedia – French – Word Count – 4 month range
Mapping Wikipedia – Arabic – Word Count – 4 month range
