Just saw that Google Correlate has gone live, analyzing search activity patterns and real world trends. Already hooked after a couple quick tests. Disclaimer: this is not serious data analysis by any means, just playing around.
Finding correlation between real world trends (e.g. winter, seasons) and search terms is the main point. But it’s interesting to see correlations just between search terms themselves. For example, the search activity of both “depression” and “magnetism” are strongly correlated. Both also seemed to dip with winter/summer in 2010:
Zooming out to 2004-2010, that winter-summer-winter cycle definitely seems to repeat itself in this chart generated by “depressed” vs. “why”.
Noticed the sporadic drops at the end of every December. Searches for “happy” are the opposite. Couldn’t resist posting the chart with the correlation between “happy” and “chanel wallet.” (Update: Of course correlation does not necessarily mean causation, though it’s a juicier conclusion. It’s much more likely that both “happy” and “chanel wallet” are simply both subject to the same external variables.)
The drawing tool is a lo-fi touch. Without uploading any real data, rough drawn lines can stumble onto interesting results. I tried my best to draw a straight descending line from 2004-present. Some results: “Phone books,” “flash web,” “telephone numbers.”
Update: Was curious about the re-popularizing of old electronic instruments like the Roland 808 drum machine invented in the early 80′s. Looks like Son Goku (a Dragon Ball character?) went through a similar resurgence pattern, with the top correlation spot. (I’m sure the sudden 808 spike in 2008 is due to Kanye West’s album.)
Casio keyboards and Braun Shavers? Looks like they’re 1. Both good holiday gifts maybe? and 2. Were both slowly losing appeal but regained interest in 2009-2010.