I was looking for a simple easy way of normalizing the distribution/histogram wrt to the sample pts.
We have response Y which is measured as a function sample pts (N). There also exists a group for each sample, available from 2 total groups.
The total sample sizes of each group being quite different from each other distorts a historgram where the two group are a plotted together wrt to bins.
Is there a direct easy way to convert the actual numbers to a density function, where
count in any given bin wrt to group is divided by the total number of samples in that group
histogram would then show the densities of each bin.
We could go on computing the required data to then plot it .... I figured there may be an easier way ....
It seems to me that if you have the number of points per bin, that you could use that as the "Weight" variable, and get the results you desire.