I took a stab at it, but I fear I didn't grasp the full scope of the issues at hand.
A few questions:
Are you seeking to keep the color ranges continuous, or would a range of say 5 discrete colors suffice?
Do you know in advance which metrics have skewed values, or would you have to figure that out on the fly?
Basically, all the attached is doing is just splitting the Selected Metric into (in this example) 5 groups
based on the range of the Metric. There are three different examples of splitting the metric: one
where there is more separation at the low end, one were there is more separation at the ends,
one where it is evenly spaced. More splittings can be made, if you know what the distribution looks like in advance.
And then based on the parameter selected, one of the splits is chosen.
204822col.twbx 681.6 KB
There is one more approach to the negate the skewness that you can try -- Using percentages.
For example : If you are looking for male, female population as a measure on Color shelf ( which can be very different for different counties ), What you can try is male/female population as a %age of Total population .. This might unskew your data and you will see a better range of Colors.
Thanks for reply...sorry I am little late in getting back to you. We got a different project last week and today is the first chance I had to get back to this.
I would like to use 5 discrete colors like in my previous screencap, and then distribution on the map would change depending on the range for whatever metric is selected. When the user drills down from region level to individual county, the color ramp will automatically update for the values that belong just to that county. When they zoom back out to the region level, the color range resets to the region's range of values for that metric. Right now I have a 3-color diverging ramp set to 5-stepped color.
I can get the ranges for each metric in advanced, if that helps.
Is there a way for me to privately share my workbook with you? I'm not permitted to share it publicly d/t work policy, but I can share it privately.
Thanks for your reply, too! We are displaying the absolute values, percent difference, and percent change at the municipal level. I'm not sure if we could use percent of total for all of our metrics, but it might work for one or two. I will give it a shot and see how it looks.
Hmm, changing the scale with drill down wouldn't work in my scheme.
Will need to think about it more.
I think I'm getting the gist of what you're after, but if you'd like,
it may be easiest if you could post just a small set of made-up data
that looks similar to your actual set.
1 of 1 people found this helpful
Apologies for the delay and meager progress.
Was looking at this page describing graphing skewed data
Tossed around some ideas about dual axes with transparent colors.
Here, the colors don't mean that much but at least it shows some separation.
Also tried just a logarithmic scale. Again, harder to grasp the color of the log of the metric.
(The dual axes were of a straight scale over a logarithmic scale).
Will continue to try some more.
Hi Swaroop, it is many months later after our initial efforts, but I wanted to let you know that we eventually got this to work using your example. We ended up using nested IF statements within a Case statement and hard-coded the ranges we wanted to use, after obtaining the data ranges for each measure in ArcMap. It worked with the parameter and we were able to show the data ranges in the legend. Maybe a little clunky with the work-around, but hey, it worked, much to my relief! Thank you very much for your help.
Glad to be of assistance.
Thanks for the feedback.
All the best.