A colleague recently pointed at this chart from Google Trends showing searches for 'recession'. It's interesting to see this relatively technical economic term catch the zeitgeist in real time. But what he was really showing me was the way searches for recession rose sharply significantly in advance of Monday's financial markets bloodbath. His supposition of course was that you can use these data to predict the market and stay one step ahead.Now, obviously this single data point does not incontrovertible evidence make, but I'd be willing to bet that the aggregate query stream is full of insights that could help you predict short term future. To that end I think that this initiative which google.org is funding is especially exciting. Essentially they want to use all sorts of analysis of data (think news sources, query streams, blogs, anything you like really) to try to predict rising epidemics before they get too big to contain. I'd trade the ability to spot the next recession for that sort of capability any day.
That said, I'm interested what people think we can predict using these data and more importantly what keyword I should be tracking to spot the bottom of the downturn!
2 comments:
Good Post. I suggest reading Wisdom of Crowds. In the bit about the Columbia Shuttle disaster Surowiecki points out that financial markets determined which part of the equipment was faulty before the investigations did, by penalizing the supplier company's stock price.
Thanks Omid. I read wisdom of crowds and liked it. My only concern was that the anecdotal evidence (like challenger) is nice, but unless an approach (e.g. the stock market) consistently reveals this info, it's not that useful. Hind sight is a wonderful thing (and lets you pick examples to back up your central hypothesis!)
A good read on that topic is 'fooled by randomness' you may want to give it a go...