The data size was adequate to generate some sort of meaningful insights but a larger sample size could have utilized to a larger degree the capability of the GWR operation. In my case across the US, there weren’t many Microgrids and missing data in some cases have contracted the sample size even further.
GWR is most reliable with large sample size.
The figure above highlights regions where there is an association between income and the amount of Microgrid employed. Areas with a lighter shade indicate regions where there is a weak association between age and microgrid deployment, and the degree of association improves with the increasing darker tone of the color palette.
The figure above helps in highlighting the areas which have an association between age and capacity of Microgrid employed. Brief description of results you obtained.For the independent variables risk index, Total population, household income, Median age were taken as independent variables. The layer was then used to compute our geographically weighted regressions.įor the Geographically Weighted regression the Total capacity of the regions were taken as the dependent variable. The data thus formed for the risk index in each microgrid position was then joined with the spatial join function to the FEMA disaster risk layer data with the help of the common county FIDs for each of our points. This data for the individual microgrid was enriched with the “enrich” tool to incorporate census data about total population, income, age of the regions in question. The microgrids were overlayed in the disaster vulnerability layer and for each point a buffer was drawn which allowd for a mean index of the disaster risk to be calculated for each point where the Microgrids were positioned. To understand how the spatial positioning of Microgrids are related to disaster vulnerability and other demographic factors, a Geographically weighted regression was run. Maps of Geographically Weighted regression would help in determining the discrepancies in these logical conclusions. Higher income and microgrid capacity would also have a direct association. I have hypothesized that regions with higher disaster risk would have a higher capacity of Microgrids installed, regions with higher income would have higher Microgrid capacity, with increasing age the capacity of microgrid available in a region would decrease, with higher population there is a greater likelihood of having higher Microgrid capacity. The Questions that were tried to be addressed were whether the capacity of Microgrids in a particular region is dependent on the disaster risk index of the region, and some other demographic features of the reason.