Hi, @RandomEtc @mbostock, excuse me, but I don't like the default extent of cubism.
If here is a much bigger number in the series, for instance, a series with average 10 but maximum 1k, like this chart:
The default extent is using max and min, but this makes the chart unreadable.
I know that we can set a const number (e.g. 0, 20) to solve this. But what my situation is, there are thousands of metrics to plot, I can't set them one by one.
My solution to this is using 3-sigma rule, and the extent will be: average - 3 *stddev, average + 3 *stddev. It covers most datapoints, no singular values.
So I open this issue to ask you how I can implement this.
Hi, @RandomEtc @mbostock, excuse me, but I don't like the default extent of cubism.
If here is a much bigger number in the series, for instance, a series with average
10
but maximum1k
, like this chart:The default extent is using max and min, but this makes the chart unreadable.
I know that we can set a const number (e.g.
0, 20
) to solve this. But what my situation is, there are thousands of metrics to plot, I can't set them one by one.My solution to this is using
3-sigma
rule, and the extent will be:average - 3 *stddev, average + 3 *stddev
. It covers most datapoints, no singular values.So I open this issue to ask you how I can implement this.
Sincerely, thanks.