I'm having a question on how the data for the long term players per day is calculated.
e.g. for the adventcup qualification server, which is a TA qualification from Monday to Friday, we can see the follwoing graph:
From my memory the server was always very crowded during the first TA qualification (Nov 23 - Nov 27).
Now I'm wondering how the average player can spike so much in that graph above?
In the source code of the site I found that you plot the following data:
Code: Select all
xdata : "2015-11-23 19:50:19,2015-11-24 19:52:19,2015-11-24 10:16:20,2015-11-25 10:16:20,2015-11-27 10:16:20,2015-11-28 10:18:19,2015-11-29 23:42:19,2015-11-30 23:44:19,2015-12-01 23:44:20";
ydata: "6,8,0,0,7,2,1,1,1";
Wondering me is that there are different times for each day on which you plot the average players.
So that's why I'm asking. In which intervals do you calculate the average playernumbers per day?
To me it looks like you take relatively random time intervals, so that we happen to have (as an example) two data points for one day - One datapoint with he average playernumber over night until the morning and another one with playernumber from the morning till the evening. Now as playernumbers vary a lot over the daytimes (more players in the afternoon and evening) datapoint2 is a lot higher than datapoint 1.
I suggest to calculate the daily playernumber by an always fixed time interval. e.g. from 0:00 to 24:00 CET. Then you can actually have a better comparison of the numbers of each day
PS: For consistency, why did you pick two different graph layouts for the same page? Especially for a longer period of time with many datapoints the big dots connected with a smooth line are not a good option:
