The data set is not even manipulated. The graph in itself is not worthless, not wrong, not even uninteresting, it’s all the stuff around it (the interpretation) that’s bullshit.
You should probably never use such a skewed graph alone, though, without showing growth or decline. Relative shifts are genuinely easier to see with a graph like that (but internet traffic is probably not the right measure if you want to know something about cultural changes, time spent would probably be better; to wit, reading 1 MB is a heck of a task, listening to 1 MB won’t keep you busy for very long and blink and you will just have missed your 1 MB video).
You should probably never use such a skewed graph alone, though, without showing growth or decline. Relative shifts are genuinely easier to see with a graph like that (but internet traffic is probably not the right measure if you want to know something about cultural changes, time spent would probably be better; to wit, reading 1 MB is a heck of a task, listening to 1 MB won’t keep you busy for very long and blink and you will just have missed your 1 MB video).