A post on LOOSE wire led me, via Corante's marketing hub, to this utility for producing sparklines. I've used it to compare board mentions of YouTube.com (
) with blog mentions (
). This is fun, but it doesn't really get to a comparison. Even if we line them up vertically:
The fact that they are normalized means we can't do a real comparison of the daily values. The red lines indicate values > 90 which is part way there, but I'd like to see something with a higher data/ink ratio. I wonder what could be done to create comparative sparklines - just for two trends.
Remember that sparklines are mostly useful when the typical slope of a trend is +/- 45 degrees and we're detecting slope changes. I think you would have to take longer term averages to get more meaningful appearances, and pick that average wisely.
Posted by: Matthew Siegler | March 21, 2006 at 07:33 AM