While there is some data that is useless and potentially harmful (why do we need to know how much bacteria there is on our toilet seat?) It’s worth noting that Chronosanct is compiling this data on their own free time.
That said there are various tendencies including launch numbers decline bandwagon-jumpers falling off the bandwagon… etc.
Interesting data to speculate with later. It seems you’ve missed the peak hour by one hour. I don’t think it matters much but the peak values are systematically off, excluding the first data point.
If you correct this, it happens to increase the R^2 slightly, while the x-parameter drops by 13% to -1150x. And if you update today’s value in, R^2 increases slightly more while x-value decreases to -1024x.
But since this is the first week I don’t even bother speculating yet.
It will be interesting to see how the consecutive week days’ peaks will develop after the F2P opening rush 
According to the data, the peak activity is probably between @12:00 and @14:00
For example today (4/8/2016)
@12:00 26199
@13:00 26304
@14:00 25480
Basically, the most important thing is that I sample consistently the same time point. One could sample any point in the curve as long as it is sampled consistently (and any outliers/aberrant trends due to maintenance or login issues are tossed).