Now this is very different from the picture in @arcange's report. I suspect the reason is that @arcange is sorting the bars incorrectly (so that the tallest, "Dead fish" bar is accidentally shown as the "Whale" category).
There are no error in my chart.
It is because unlike SteemitBoard, I only kill inactive red fishes for my charts.
My main goal in this report is to represent "real" distribution among levels and the "inactive level" is just based on arbitrary criteria, even if it might be an interesting one.
By killing only red fishes, I remove only all those fake or never used accounts.
RE: DIY Steemit Statistics with Python: Part 6 - Voting Power