Abstract

Discovering transport and land use urban patterns using Geostatistics
Track: Transportation
Authors: Daniel Paez

This presentation explores factors affecting private car use for journey to work in Melbourne using geographically weighted regression (GWR) analysis. With this geostatistical methodology in ArcGIS it was possible to determine key factors determing people selection of transport modes for traveling to work. Multiple variables were explored including land size, proximity to community facilities, commercial activity and provision of transport infrastructure. The use of GWR in ArcGIS increased the explanatory power of the analysis (R2) from 57% in the case of simple regression to 74% when GWR was used. GWR also enabled spatial patterns of the major explanatory variables to be explored. For this particular case in Melbourne, public transport supply was found to be particularly strong at influencing private vehicle commuter in inner and south western parts. Distance to the city was stronger in inner and northern parts of Melbourne. Residential density was strongest in eastern and outer areas.