Skip to Content

An ArcGIS Add-in for Geographically


and Temporally Weighted Regression Analysis

Professor Huang Bo, from the Department of Geography and Resource Management, Chinese University of Hong Kong, has developed an ArcGIS Add-in for performing analysis based on Geographically and Temporally Weighted Regression (GTWR) model.

Professor Huang developed GTWR model to analyze data with spatio-temporal heterogeneity such as real property prices

Professor Huang developed GTWR model to analyze data with spatio-temporal heterogeneity such as real property prices.

There are a number of reasons to use regression analysis based on different factors including geographical location and time:

  • To model different phenomena in order to bring a better understanding to the issue and possibly make a better decision
  • To model different phenomena to predict values at other locations
  • To explore hypotheses

Regression analysis, when performing under a spatial statistical perspective, it can examine and explore the relationships between different factors and taken their spatial location into consideration. It is also possible to make predictions base on multi-dimensional parameters.

ArcGIS provides Geographically Weighted Regression (GWR) tool that generates spatially calibrated regression models. By incorporating temporal effects into the GWR model, the geographically and temporally weighted regression (GTWR) model developed by Prof. Huang can deal with both spatial and temporal nonstationarity simultaneously in various kinds of data. It integrates both temporal and spatial information in the weight matrices to capture spatial and temporal heterogeneity.

t has been proved that better accuracy can be achieved when analyzing housing prices based on GTWR model

It has been proved that better accuracy can be achieved when analyzing housing prices based on GTWR model.

With this tool, it can help to analyze the correlation between different factors affecting the property price in Hong Kong, while bringing the time element into the analysis. In fact, location and time are important determinants of real estate prices. Professor Huang is able to show that there is substantial improvement in analyzing housing prices when GTWR model is used. It is believed that the GTWR model can also provide benefits to other analyses that involve spatio-temporal heterogeneity.


The ArcGIS Add-in (GTWR Beta 1.0) can now be downloaded from here.


Please find below the links for more information about spatial statistics and regression analysis:

1. Getting Started with Spatial Statistics

2. Regression analysis basics

 

 

 

 

 

Back