Statistics for the Statistically Inclined
June 10, 2011
Due to a strong bias against everyone’s favorite search engine, it is difficult for me to become excited over new Google developments. However, having endured a number of statistics classes, I will certainly give credit where credit is due.
I was recently directed to Google Correlate and spent a solid twenty-five minutes entertaining myself with test statistical relationships. The offering consists of comparisons of an uploaded data set against a real data set courtesy of the search mogul. Google provides results based on a Pearson Correlation Coefficient (r) nearest to 1.0, giving the user the most positively correlated queries. One can customize the results in a number of manners: for negative relationships, against a time series or regional location, for a normalized sine function or a scatter plot, etc.
For any glazed over eyes out there, the Web site sums up the intent this way:
“Google Correlate is like Google Trends in reverse. With Google Trends, you type in a query and get back a series of its frequency (over time, or in each US state). With Google Correlate, you enter a data series (the target) and get back queries whose frequency follows a similar pattern.”
Don’t worry, there is a tutorial.
It should also be noted that this service is tagged as “experimental”. I fear due to lack of popularity, it may dissolve in its very own time series in sad, monthly increments.
I imagine this tool is providing certain students some relief, but what of regular users? In the words of the head gander, how many Google mobile users know what correlate means? Without crunching the data, I think our r may be approaching -1.0.
Sarah Rogers, June 10, 2011
Sponsored by ArnoldIT.com, the resource for enterprise search information and current news about data fusion
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
