Google Blogoscoped

Friday, June 13, 2003

Googleshare Categorizer

The Googleshare competition based Categorizer went online. It’s working similar to the Centuryshare calculator. Note that the Google Web API access seems to be unstable at the moment.

The Categorizer has 12 categories (like travel, place, person, clothing, ...) and will try to put any word or phrase you enter into one specific category. Once the Google API access is working stable again I will try to introduce levels of sub-categories (like: place - building | country | city | planet, or art - impressionism | expressionism | cubism, etc.) to further narrow down what the query is about.
The core-script is only a few lines long and yet it worked remarkably well; e.g. “shoes” were put into the “clothing"-category, “nihilism” into “concept”, and so on.

End of Google Premium Sponsorships

“Today, Google VP of Advertising Sales Tim Armstrong made it official: no new Premium Sponsorships will be accepted and existing sponsors will not be able to renew at the end of their terms. This means that the program should be history by year end.”
– Andrew Goodman, Confirmed: Google to Phase Out CPM-Based “Premium Sponsorships”, June 12, 2003

I’m currently experiencing problems accessing the Google Web API with all of my tools. It can’t be my Google API key really, since my friend with her own key has the same problems. I wrote to Google support and will post updates here.

New Centuryshare

The following is a chart for “NASA” created using the Centuryshare Calculator’s new optional level of high detail (covering 40 years of the 20th Century, year by year):

An early peak is the creation of the NASA (the National Aeronautics and Space Administration) in 1958. The highest peak for all years is the Apollo 11 Moon landing on July 20, 1969. The highest peak of the 1980s is the tragic Challenger explosion on January 28, 1986.

I’m quite sure that with a more or less unlimited number of available Google Web API requests, one could continue the approach and calculate a date for the peak. The queries would then go like 1969 nasa january, 1969 nasa february, 1969 nasa ... and so on. After finding out the highest page-count month, one could then continue with 1969 nasa “july 1”, 1969 nasa “july 2”, 1969 nasa ..., continuing up to 31. A calculation would be very slow then, as it would take 40+12+31=83 requests (provided the data for the average page-count for every month, and every day in it, for every year would already be stored by the program, making for almost 40,000 values — which of course would have to be pre-requested automatically).
But the result would be interesting, as it would show the “most important” date in history for any given keyword. E.g. entering “birthday einstein” would most certainly result in the correct birthday for Einstein.

One more time, in comparison, the Beatles (1940-1980):

Punk (1950-1990):

I don’t really know why “Punk” has the same peak as “Beatles”, 1963. Either something is going wrong, or Punk was heavily influenced by the Beatles (possibly as counter-movement to earlier Rock ’n’ Roll, like progressive-rock, concept albums like Sgt. Pepper’s, and art rock).

And Glam Rock:

Here’s Stanley Kubrick and his films (2001: A Space Odyssey ranks highest):


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


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