Google Data Mining

If you wander about the Internet long enough, connections appear, sometimes answers to questions. Last fall I wondered about what Google is going to do with all the databases and information it collects on all the users of its various businesses. Also noted was addition of a sort of ‘zine to Google’s social networking the eternally-in-beta, Orkut. Keeping an eye on Orkut Media paid off. An ostensible contributer, Ellen Spertus, is dropping in interesting little tidbits about the vast interconnections among Orkut members. Although she doesn’t mention how she comes by the data, it’s obvious she’s either a Google plant or someone who’s acquired access to the data. A blurb on her at the bottom of her articles states that she “moonlights at Google”, whatever that means.

She is a member of Orkut, and started her column on Orkut data in January with an article about the groups the system recommends (like Amazon.com’s book recommendations–If you like this, you might like that.) Pretty innocuous stuff. In February she asked, “Which orkut users most successfully obey the request to “stay beautiful”? This column takes you into the data mine to find out.” My ears perked up. Data mine? She began giving percentages of members who describe themselves a certain way, and then she broke it down by country. She also started a new Orkut group, Too Much Information, “For discussion of internal Orkut data mining, as described in occasional Orkut media columns. Comment on past research or propose new research here.”

In this month’s offering Spertus digs deeper into the personal ranking system that Orkut calls “karma”, breaking out some features by ethnicity. She says she’ll give us more next month and warns, “Don’t try to game the system, though. I’ve already extracted the numbers, and new measurements won’t affect them.” Sounds like more than moonlighting to me.

Send your comments to Georganna at Writer’s Edge.

Article Written by