I would disagree with that to an extent for the distribution below "I excluded the 20 most repetitive songs in the dataset from the above chart. When you see what it looks like with them, you'll understand why". I had both distributions on screen so I could compare them, and was staring at them for a good while thinking I was missing something obvious. The animation that actually revealed the difference didn't start until I'd given up and scrolled the diagram most of the way off the screen. Had I scrolled a little faster to the next section (or possibly used page down/space), I would have missed the animation entirely.