Hello,
I am working to get a time series trend. I have data over a 30 year stretch for fixed polygon locations. I'm not really trying to look at anything spatially just yet, I've just been looking at getting a trend value for each polygon and then plan to look at what spatial patterns there may be.
I have a model built with the OLS tool where I am splitting up polygons to run the OLS on each set (i.e. my stationary time series data) and then appending the stats parameters to an output table to join back to the spatial extent. It works, but it is not at all an elegant process. I will need to run this analysis again and hoping to simplify.
It seems that the GWR tool should be able to give me a coefficient value for each polygon if I ask it to consider the local area for the calculation as the polygon extent for my stationary data. I've tried this using a small fixed bandwidth, however it will not work.
Any suggestions for a better work flow for this? Or just a good alternative?
