Saturday 25 December 2010

Treasure troves of new data

Here are two interesting projects that allow us to significantly improve our ability to measure important economic and social phenomena. The first, provides us with a way to calculate daily inflation rates for several countries by piecing together billions of prices from online retailers.While this is for now mostly confined to developed countries, the project is rapidly expanding to the developing world, where official inflation data tend to be less reliable.
The second is a hot off the press tool developed by Google to mine digitized text for word usage. Taking advantage of Google's massive efforts to digitize over 15 million books, researchers created a user-friendly tool that counts how frequently strings of up to 5 words appear in books in different languages (German, Spanish, American English, British English, etc). This is a good start and reveals some interesting patterns. We look forward to future iterations of this product, especially once it starts including searches through other forms of mass media, which are perhaps even more norm and culture-defining than books.
Happy Holidays!

No comments:

Post a Comment