2

I have a map of the dependent variable, median household vehicle travel, and independent household characteristic variables, median income, mean household size and median age, aggregated to census areas. I'm looking to find a way of controlling for these demographic variables and produce a map of the vehicle travel assuming equal household properties across a city. The goal is to estimate how much travel a hypothetical household would carry out if they moved to different areas of a city.

Using spatial statistics and the geographically weighted regression tool I have produced a regression equation relating these variables to travel with an R2 value of ~0.5. The resulting residuals are spatially autocorrelated which is what I want as it shows some missing explanatory variable is influencing travel. It is the influence of this unexplained variable that I need to map, however, simply mapping the residual doesn't really tell me how much travel an average household would carry out, only that certain areas travel more than or less than what is expected from the regression off household characteristics.

I don't have much stats knowledge, but I'm guessing there is a way to do this. Could I use the resulting regression equation, R2 value or predicted and observed travel values to estimate the travel of an area assuming average household characteristics? Would something along the lines of the following work:

LocalR2*(Intercept+Coefficent1*AverageHHSize+Coefficent2*AverageIncome)+(1-LocalR2)*ObservedTravel 

Or will I need to delve deeper into the statistical methods.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.