When you can provide some accurate data that isn't skewed by grouping togther raids that were available for a year with those wich were available for a month, when you have data that doesn't group together data gathered using very different methodologies, when you have dtaa that is accurate, uyp to date and analysed then you can spout the resulting graph.
This is not evidence of anything except you think it important. It makes no effort to account for different data collection methods, different timeframes, other activites such as levelling a character, it doesn't take into account LFR and so on. You don't mention if the data sets are comparable.
All you've really done is shown us a set of numbers, stated they are from WoWprogress and posted them again and again and again. If you are going to throw facts to prove what everyone knows and accepts - that 25s are declining in numbers and popularity - they'll need to be a little bit more concrete than what you have given in order for us to be able to draw any conclusions other than what we already know.
What your graph DOES show is that the two sets of figures for LK show about 2.3 million each. Relatively steady. Relatively stable.
What it also shows is a figure of approx 700k for the raids of Ctaclysm. Also relatively stable.
What it does not show is why the figures dropped between the two groupings. That's actually kinda important.
Did they drop because the game losts subs? Did they drop because WoWprogress changed its data mining process? Did they drop because what used to have two achievements now only had 1? Did they change because, with shared lockouts, players now only raided one format? Do these figures account for Normals and Heroics? How many players went to LFR?
And you'd be wrong. Raiding always been a minority interest in WoW. If the 2 million who did quit had much interest in it, there'd be no raiding scene at all. Doubtless, a lot of those who did leave did raid to one degree or another. The error is in assuming the raid model is the prime/sole/major factor.I'd say a very large portion of the players that quit, DID care about raiding.
In essence, your graph tells us something we already knew. It doesn't tell us anything we need to know.
As it is, its still fairly healthy in many ways. The population just got skewed, for whatever reason, to 10s. And thats only wrong to whatever degree that 25s are forced into 10s.