BRAIN TEASER: Spatial Geography Optimization

[jmal]

I have a data set that contains a list of 100 cities within the US (Target). I also have a list of all the zip codes in the US with they're associated populations (Universe). With that information, I've managed to calculate the total population within 100 miles of all the cities.

However, if I am limited to choosing only 10 cities in my target list, how do I know which 10 cities to pick to maximize population reach? Meaning, is there a way to calculate an optimized model that favors trade areas with the highest populations, trade areas with the least overlap, etc.?

[cplewis90]

Hey @jmal,

I am trying to think about the least overlap, but as far as the most populous trade areas, you could just calc all then sort by population and once complete, sample the top 10 records. This would always give you the top 10- most populous trade areas out of the 100. 

(Initial thought) As for the least overlap you would have to find the area of the remaining non-overlapped trade area and then do the calc of pi*r*r - (calculated trade area space). This would give you how much is removed from the original trade area and you would always want the closest to 0 of that calculation. I know you could probably run this through the optimization tool so it would get you the most populous/least cut off, but I would have to think through that some more. That tool is quite tough to wrap my head around. I feel it is a new tool each time I use it.

[amanda_payne]

Have you checked out this alteryx Advanced Certification Prep video? Advanced Certification Prep - Part I - Alteryx Community