Using quantitative analysis (somewhat similar to Google's pagerank) on the english-language Wikipedia's entries to rate historical personalities, the authors come up with a quantifiable measure of their significance. This is an interesting undertaking in and of itself, and while such data mining isn't a replacement for traditional social science (and neither do the authors suggest that), it can be a very useful tool to complement more traditional assessments of the influence/importance of historical figures.

What is interesting is that even though one would expect such an analysis to inherit the biases and blindspots of the dataset they are working with, it actually highlights some of those shortcomings. For instance the authors found that of all the men and women mentioned in Wikipedia, the average significance of women (even though there were far fewer of them) was higher than that of men, implying that the threshold below which people are considered insignificant to warrant encyclopedia entries, is much higher for women than men.

It would be interesting to see the same analysis extended to other large datasets which would give insight into who is considered significant by other cultures and groups of people.