A few random thoughts:
If you measure not incidents per year, but victims per year, the darker colors would have much more impact. I'm not sure how it would change the peaks and troughs of the graph, but it makes me curious.
I don't think trends in something with so few numbers are helpful for assessing much of anything.
I was ready to say that the trend of lower incidents correlated with the last 30 year's tendency toward more gun control. This would be some evidence that gun control works (some evidence /= knockdown argument). But when I look at whether we actually have stricter gun control, I'm a little bit surprised that we may not (here first, then here and here). So I have to reverse the argument and admit that this is at least some evidence that gun control does not work.
