Wikipedia is a fascinating resource and the most famous example of crowdsourcing — on the site, hundreds of thousands of people come together to help us understand everything from Justin Bieber to hay fever.

The value of Wikipedia articles is usually considered to be in the words and explanations that users contribute to the Wikipedia pages. Kalev Leetaru of the Unidersity of Illinois and the folks at SGI, however, decided to analyze the site’s data in a different way. Loading all four million English-language pages into the SGI UV 2000 — dubbed the “Big Brain computer” — Leetaru identified the locations and dates mentioned in the articles and then analyzed the connections between them.

The computer found 80 million locations and 42 million dates in the articles. Analyzing the data in near-real time, Leetaru created algorithms that were able to visualize not only how the locations connected to each other through time, but also the “tone” of the articles. The algorithm measuring the sentiment appears to match the real world understanding of events, though it was perhaps skewed by the fact that only English-language articles were analyzed; the American Civil War topped World Wars I and II as the most negative time period in the last 1,000 years.

Leetaru also came up with several graphics depicting the connectivity structures of the data, as well as charts showing some interesting trends. One of the more intriguing charts plots the total number of articles written between 2001 and 2011, and the total number of mentions of dates of the same year. The graph suggests that more early Wikipedia articles were written about current events than today, implying that editors are now going back in time to fill in the gaps in articles about past events.

To read more about the project and to see more data visualization, check out the project page. We hope that more experiments like this will be conducted in the future for other languages so we can better understand how non-English speakers view certain events.