Bad Graph Interpretation of the Day: Well, if I had to guess, I would say bad math happened.

Saw this tweet this morning with 2.3M views:

Just eyeballing it, from 1995-2000, boy suicide rates fell from about 18 to around 9, a 50% drop. Girl suicide rates fell from about 4 to about 2 per 100K, a 50% decline. Correspondingly, the overall suicide rates fell from around 11 to about 5.5 per 100K, a 50% decline. 

I'm not seeing the claimed gender difference in the fall in rates.

I think the more relevant question is why the strong rebound from 2007 or so on? 

I'm going to posit that the rise in social media might play a role.

But that's just a guess.