# How to create weighted sum based on variety of features, rather than number of a single feature? [closed]

I am interested in creating a weighted sum analysis of three different feature types in my geodatabase. These features are services in the community. Two of the features have much greater prevalence than the third feature. I would like to create an analysis that determines the best proximate location to all three feature layers rather than a straight total that could over-value areas of high instances of a single layer.

For example, in the below area I'd like the red circled locations to be a higher weight than the blue circled locations because there are green, yellow, and red features in the red circle areas rather than just a lot of yellow features or yellow and red features:

Obviously Weighted Sum on its own would be insufficient because even if I gave each feature a value of 33%, for example, if the yellow feature had a value of 3, but green and red were each 0, it would be the same value as a location where green, yellow, and red were each 1. Is there a way to have an exponentially lower increasing value the more of one feature exists (e.g., 1=1, 2=1.5, 3=1.75, 4=1.87, etc)?

What tool or expression for Weighted Sum could best accomplish this? I have access to both ArcGIS and qGIS.

• The polynomial y = 0.0229x^2 + 0.2655x + 0.7114 fits the values you provide. – reevesii Aug 19 '16 at 20:36
• This question might help:[gis.stackexchange.com/questions/86866/… – reevesii Aug 19 '16 at 21:01
• As it stands I think this is too broad because by asking for a solution in two desktop GIS products you are effectively going against the Tour where it says "one question per question". – PolyGeo Sep 24 '16 at 6:03