The graph tracks blog mentions of Paul Krugman, Thomas Friedman, and David Brooks over the past month.
This is a really neat graph. What is even more neat is that you can make your own graphs for any keyword!
Are these graphs statistically significant? It seems not especially meaningful to use "percent of all blog posts" as the Y axis for this particular graph. You need more data to make sense of this.
There is definitely a trend in this graph, but it isn't necessarily the one that is implied by the title of the post.
It could be that the absolute number of Friedman posts has stayed exactly the same over the past month. Instead, we've seen an incremental growth in the total number of posts.
If you do a least squares fit to the Friedman line, it looks like there's been about a 50% drop in the percentage. In other words, in early September, Friedman was mentioned in approximately 6 out of every 100,000 blog posts In early October, it's down to 3 out of every 100k posts.
What this says to me is that Friedman was a blip in the blogosphere before, and he is a blip now.
Look at terms like katrina and rita over the last several months. Now there's a meaningful graph. Or consider mentions about the SCOTUS nominees.
I could make a chart comparing louis armstrong and thomas friedman, and if you look at the period from 3 September to 3 October, you could just as well talk of "the waning influence" of Louis Armstrong.
If you compare johnny cash and thomas friedman over the last two months, you'll find that Thomas Friedman has about as much "influence" on the blogosphere as the late Johnny Cash. Interestingly, you'll also find that a curiously large percentage of the peaks and valleys coincide on the two graphs, suggesting that other factors are at work. The same can be said for sheryl crow. In other words, garbage in, garbage out. This isn't exactly a well-groomed data set.