Googleshare

Steven Johnson had an interesting idea for using Google to calculate mindshare for a particular concept. You search for the concept, and note how many pages it returns. Then you search for the concept plus a person/company/whatever, and note how many pages that returns. Divide the second by the first, and that's your Googleshare. For example:

  • programming: 32,700,000 pages
  • programming "programming languages": 3,430,000 pages

From which we can deduce that programming languages have 10% mindshare within the field of programming. Now let's see the mindshare within "programming languages":

  • "programming languages": 3,430,000 pages
  • "programming languages" Java: 1,510,000 pages (44%)
  • "programming languages" C++: 975,000 pages (28%)
  • "programming languages" Perl: 787,800 pages (23%)
  • "programming languages" Scheme: 228,000 pages (6.6%)
  • "programming languages" Prolog: 94,600 pages (2.8%)
  • "programming languages" Haskell: 60,200 pages (1.8%)
  • "programming languages" (Kernighan OR Ritchie): 22,100 pages (0.6%)
  • "programming languages" Peyton Jones: 11,700 pages (0.3%)

Out of curiosity, I tried comparing the googleshare of men versus women, and I was surprised to find that women have a larger googleshare for politics, programming, war, and sports (all stereotypically masculine fields). In fact, it's difficult to find any area where "men" has a higher googleshare.

Posted on November 13, 2003 04:52 PM
More blog articles

Comments

I do this with google news. Its a good way of finding out how much news coverage something is getting. Its also surprising sometimes to toss in something odd like "Buffy" or a band or book you like and see how far in the cultural pool it has spread.
I guess while what you are talking about is mindshare, I am talking about the 'Q' rating of something.

Posted by: drlloyd11 at November 13, 2003 06:00 PM

I know these sites are probably old news, but googleshare reminds me of Googlewhacking and the Google Zeitgeist . It's fascinating to see how Google has changed Internet usage; not only does it dazzle and entertain, but its logs are apparently becoming valuable social reflection.

Posted by: Michael Tucker at November 13, 2003 09:12 PM

While I haven't looked into this, I'd guess women beat men here because they're the "marked" case -- i.e., when men are implicitly taken as prototypical they aren't named explicitly.

Posted by: Darius at November 14, 2003 02:14 AM

Alex Martelli has been playing with this concept for a while. See:

http://groups.google.com/groups?selm=HJrob.386734%24R32.12803775%40news2.tin.it

for a recent thread on the subject.

Posted by: Simon Brunning at November 14, 2003 04:46 AM

Let this be a lesson to you: when you invent the successor to Haskell, be sure to give it an obscure name so it will stand out in searches.
For example, a google on
bioinformatics "Common Lisp"
or
bioinformatics OCaml
is much more useful than
bioinformatics scheme
I didn't even bother trying it with C...

Posted by: Ralph Richard Cook at November 14, 2003 05:05 PM

Just saw something like this applied to the U.S. presidential candidates: http://www.googlerace.com/. An interesting use of the Google API...

Posted by: Sanjay at November 22, 2003 01:32 PM

Women use computers more than men. And this is not a sexist remark. They just seem to have more time.

Posted by: gurugarzah at April 29, 2006 04:21 AM
Post a comment









Remember info?




Prove you're human. Type "human":