In my mind the benefit of using Suitability Modeler is that you get a continuous surface normally over the entire study area showing you variations over that surface area, however at my job we wanted to do the concept of the suitability modeler without using raster data and only using vector data. We built a Model Builder project that mimicked the Map Algebra using in suitability Modeler but with the vector data. So, you only mention Unions and Intersections, well what we did was if items intersected, then we would add the attribute values for those that intersected (you don't have to add, you can use any math formula on your attributes that make sense). The other thing you need to do to mimic Suitability Modeler is make sure all your attributes are in the same scale otherwise one variable will over power another variable if there numerical scales are drastically different. So, you can use Normalization or z-Scale Standardization of the variable to get them in the same scale. Lastly, the beauty of doing this with vector data and with your own model is that you can add as many columns to your dataset for attributes instead of having to create a raster for each variable to do the map algebra because in model builder you can use values from any of the columns in your math formula making the process much simpler.
We personally used this Suitability Modeler concept with your vector data to prioritize where we go next with our engineering buildouts based on a lot of variables like Route Miles in a Community, number of households in a Community, which communities has the closest fiber to them, how many businesses would we pass on the way to build out the community, etc.... But it boils down to we only do the math of the suitability if we have an intersections (or some other type of spatial join) of the vector data. So, I believe you could do your project with Vector data if that is what you have and get similar concept of what Suitability Modeler accomplishes, only you end up with Vector data in the end instead of a raster image. So, your different vectors would have a single discrete value that you could map that is the result of your math operators you applied meaning the entire polygon or line would have the same value, where as in Suitability Modeler, that area of a polygon would have varying values of a continuous surface.
Robert “Edward” Gause, GISP | Director of Information Services | HTC | p 843-369-8483 | www.htcinc.net | This is life. Connet with it.