As the article points out there are trends and events that can be predicted with some probabilities. An interesting feature of the predictions is "self-fulfilling prophesy". The predictions themselves may affect some events or timelines.

“Lugeja” is correct about both computer programmes and the effect called “self fulfilling prophesies”. Both a simplistic computer model (programme) as well as an “over specified” one will, in general, fail in forward forecasts. The other side of the coin to “self fullfilling prophesies” is averted-predictions. Both are results of the human element in our society’s feed-forward and feed-back loops… destabilisation and stabilisation respectively.

Data Mining is now “big business” so I am not surprised to see Academia (and Leetaru) pursuing this line. I do wish him success, but am inherently sceptical… the best “pattern recogniser” is a human. The best intelligence analyst is a human. I don’t discount a supercomputer possibly becoming an indispensable aid, though.