There were a good article criticising Google Flu approach: “ The parable of Google Flu:  Traps in Big Data Analysis” http://dash.harvard.edu/bitstream/handle/1/12016836/The%20Parable%20of%20Google%20Flu%20(WP-Final).pdf?sequence=1

It seems too easy to predict anything using Google or Wikipedia data. Just follow the following plan: a) choose what to predict, b) find searches that give you a positive results.

And this unscientific approach produce articles like this one: “Predicting Recessions in Real-Time: Mining Google Trends and Electronic Payments Data for Clues.”  http://www.cdhowe.org/pdf/Commentary_387.pdf    

Thank you James, I am not criticising the article you mentioned.
I just ready to to bet that I will be able to predict anything on Wikipedia pageviews data.
-- 
Thank you.

Alex Druk
alex.druk@gmail.com
wikipediatrends.com