I just read this neat paper by Hal Varian, Chief Economist @ Google and emeritus Professor of UC Berkeley’s School of Information. Titled “Predicting the Present with Google Trends,” he goes on to explain the possibilities!
Reading his paper on predicting the present with Google Trends came as no surprise. There is significant value in accurately measuring the pulse of the present, especially for economic measures like alleviating or even bracing for an upcoming recession. His statistical methods were sound, simple, and straight-forward. I’m pretty happy that I am taking Stats 201 on Statistical Methods for Data Analysis because I was able to follow Varian’s methodology and understand the R output. His examples on the auto-industry and travel destinations were good examples, but I wonder what we can truly do with such information. In a sense, it’s almost finding out what the future will hold for us in the next 4-6 weeks. Now we are presented with the options of accepting the inevitable or fighting it to change course.
Moreover, I want to see whether social networks like Facebook, Twitter, and Google+ can contribute anything. I’ve seen talks on how the choice of words used by the media and Twitter can indicate the general mood of the population, and therefore be used in the stock markets. Just by looking at the References of the paper, many people have tried predicting various things like monitoring influenza and other diseases.
When I was interning at Twitter this summer, there was an ongoing project to measure the political mood of the US population to predict which candidate will win the US Candidacy. The method was to check the tweets using various algorithms to decide how well Mitt Romney and Barack Obama was doing, and the public’s perception of them. Furthermore, during the Presidential debates, CNN displayed live Twitter counts by the minute. On a microcosm scale, I see the entire effort of predicting the present equivalent to someone getting immediate feedback on one’s status, whether it is working out, productivity at work, and so forth! Then one can make decisions based on this new information, such as whether how sedentary that individual will be in for the rest of the day. If I knew I was going to be seated for the next few hours, I would definitely change that behavior.