Google Blogoscoped

Tuesday, June 5, 2007

Googleshare Floors

Googleshare calculates the likeliness that two things belong together or are strongly related, by looking at the Google page count for specific search queries. I wanted to try this with floors of a building this time. In order to do so, we first need to find the base page count for every floor, so we query Google for the page count for ["ground floor"], ["1st floor"], ["2nd floor"] and so on by doing some screenscraping (this turns out to be more reliable than the Google SOAP API). Then, we query Google for e.g. ["1st floor” restaurant], ["2nd floor” restaurant] and so on, and use the percentage of this new page count in relation to the larger base page count – this will be our normalized Googleshare (normalized meaning it doesn’t matter whether or not, in general, “1st floor” is talked about much more on the web than say “20th floor”).

To cross-check if we’re getting meaningful results, I’ve used the term “front-desk” (which returns the ground floor) and “suite” (which returns high percentages for almost every floor, the 1st floor being the best match). Below on the other hand you can see the results for the words “restaurant” and “suicide”:

Looks like most people start jumping from 5th floor (it probably doesn’t make much sense to try from much further down), with the 13th floor being a peak... perhaps because this number is special to the superstitious (other terms had higher percentages for the 13th floor as well).

Also see more Googleshare visualizations.


Blog  |  Forum     more >> Archive | Feed | Google's blogs | About


This site unofficially covers Google™ and more with some rights reserved. Join our forum!