domingo, marzo 19, 2006

CenSEARCHip

CenSEARCHip: "
A neat project at the Indiana School of Infomatics compares Google US and China for the same search. Fascinating to see the results build on the fly. The site uses a variation of a tag map from the results, as opposed to just the results themselves. From the about page:

When you click the 'Web Search' button, each side of the display will first show you an estimate of how many English-language results the search engine has for that national version. Our system will then begin downloading the top few pages that are unique to that country's results. As the pages are downloaded, you'll see a set of words of varying size in each half of the display.

We get those words by breaking the pages up into individual terms, throwing out some common noise words ('and', 'the', etc.), and tallying up the results. We then find the 50 words that have the highest relative frequency of use on each side and draw them in a font size proportional to their frequency. For example, if you see that the word violin is very large on the Chinese side of the display, that means that the pages unique to the Chinese search results use the word violin much more often than the pages unique to the United States search results.

You can also see image searches. Unfortunately it does not support inline URL searches so I can't link to specific sea"

Link