Also keep in mind a decent chunk of the adult content was literally automated bot spam.
Meanwhile, Statista's graph appears to be made of random numbers with no basis in reality whatsoever. They claim that even post-adult-content-ban, Tumblr has more MAUs than Twitter and Snapchat combined. Factor in Tumblr's 70% drop vs historic peak usage, and you'd end up with a number claiming Tumblr was once the most popular tech product on earth. This is quite obviously a complete fabrication.
If you want to be sad or angry about something, the loss of old peak Tumblr -- when it combined the best aspects of Twitter, Instagram, Pinterest, and Reddit -- is a better target in my mind. Its decline was long and slow over many years.
(Disclosure: I was one of Tumblr's first engineers, and later director of platform engineering. I had no involvement whatsoever in the adult content ban, but nonetheless find it frustrating when people significantly overstate the amount of adult content on Tumblr. At its peak it contained an extremely diverse set of communities, and I fear this is being completely lost in revisionist rhetoric.)