Interesting graph. The top one shows searches for "djembe"; the bottom one shows references in Google news stories.
Note that the frequency of searches has gone down, but the news mentions have gone up. Seems like conflicting trends.
Anecdotally, at least in Australia, I'd say that popularity has waned somewhat over the past eight years. There are fewer drum circles, for example, and fewer people attending those that are left.
On the other hand, there are trend indicators that seem to say the opposite. There are more teachers around now than eight years ago, there are far more shops that sell djembe-related stuff, and there are more Australian performance groups as well.
In some sense, I think the industry has reached some degree of maturity during this time. When I first started drumming, there was almost nothing available in terms of teaching materials whereas, today, there is a ton of it. Similarly, buying djembes, dunduns, and accessories was extremely difficult back then whereas, today, I can get pretty much anything I want; if not locally, I can easily get by mail order, and for better prices than eight years ago. Again, that indicates a trend up; all these ventures need customers to sustain themselves.
In terms of raw numbers of people who play, I'd say that there are fewer now. But that's in Australia only. Look at China, Japan, or Korea, and you'll get the exact opposite picture. Especially in China, growth apparently is phenomenal. Of course, when you are starting from a base of zero, there is only one direction for the trend to go. But, from what people have told me, djembe is taking off there in leaps and bounds, with huge growth rates.
I also suspect that djembe playing is subject to trends and fads, much like the skateboard craze in early seventies (substitute fad of choice here). There will be waves up and down as time goes by.