Data Analysis and Related Applications, Volume 1

19Geographically Weighted Regression for Official Land Prices and their Temporal Variation in Tokyo

This study models Tokyo official land price data using geographically weighted regression (GWR) and multi-scale GWR (MGWR) models. The GWR model spatially explores the varying relationships between land prices and the exploratory variables. Based on the estimated model parameters, the influence of land individuality increases as the estimated bandwidth parameters in the GWR model decrease. These facts are also confirmed by the local regression coefficients of the access index, the distance to the nearest station and residential area dummy variables. The differences between local coefficients for some convenience indicators, including access time to central Tokyo and walking distances to nearest stations, tend to increase between the west and central areas of Tokyo.

19.1. Introduction

In the “price announcement of public land”, published annually in March, the Ministry of Land, Infrastructure, Transport and Tourism (MLIT) of Japan announces the prices of standard land on January 1 based on the Land Price Announcement Law. This announcement not only provides an appropriate index of land prices for general land transactions, but also plays the role of institutional infrastructure in the socio-economic context, being applied to national land use plans, including land acquisition and planning for public work projects. The price announcement of public land is thus not only the standard ...