8611 wrote:In that graph, are BA & zK now not visible because they are too far up on the y-axis?
They are not visible in this graph because this graph is about "less popular" mods only, as I though you were interested in this spike particularly (it details the "Other mod" data from the previous graph). If they were drawn on the same graph, then the Y-axis scale would change and data for "less popular" mods would be far less readable.
8611 wrote:Does the "players in game by mod" graph flattens/averages too much?
What do you mean by "flattens/averages too much"?
Either you draw a raw graph like this one and personally I don't find it readable:
Either you use round robin consolidation functions which are exactly done for this purpose, to show readable graphs averaged according to the pseudo-periodicity of the data over the timespan.
8611 wrote:Of course it should not be unreadable "spikey" but on the graph it looks like there is constantly 2 XTA players or 5 TechA players.
One can easily see on the graph of my previous message that the resolution of the X-axis is one day, so you obviously can't draw conclusions about players constantly playing at all time of day. But you can draw conclusions about players playing every day. And especially, you can draw conclusions about the amount of time spent by players every day in each game, which is helpful to compare game popularity, which happens to be the topic of this thread (except as I said it's not multiplayer popularity only but global popularity...).
8611 wrote:Maybe for such data instead of "averaging"/smoothing it could be more representative to see summed-up player-minutes per week or per day?
That's exactly what my graph is showing (you just need to multiply the numbers by 24*60 if for some reason you really want to see player-minutes instead of player-days).