I spent several days there one afternoon... Ditto Booker and Lipscomb.
The crux of the problem or crust of the biscuit if you will...
The solution is to obtain a large enough data set to make some sense of some of it. And that, in a large rural area like the north half of the panhandle of Tx has to be to take in data from several counties. I do the same in NE Oklahoma by using the entire MLS. Is it perfect, nope, But it works. OTOH, a simple single linear regression works well when you have few data points if you have points that are variable enough to draw a trend line and estimate the formulas with it.