"I find it an extremely good example of the new types of data historians will get to use in the future." Potentially, but I have two concerns about this. First

"I find it an extremely good example of the new types of data historians will get to use in the future." Potentially, but I have two concerns about this. First, I'm not familiar with the facebook api, but does it does it provide a way to determine an account's like count retrospectively? Or does this kind of analysis rely on scraping pages over the time-period of the analysis? Put it another way, could a historian in twenty years time reproduce this data, or does it have to be collected now because no historically useful record is kept? I suspect that online sources actually leave much less historically useful data than, say, newspapers on microfiche, despite online data being much easier to analyse. And secondly, what happens to the data if/when facebook or twitter or google disappear or merge or change massively. As commercial entities, none of them are in the historical data business for anything other than short-term commercial gain. They have no obligation to the more distant future. I wonder what historians think about everything going digital?