2006 Volume 5 Issue 2 Pages 41-46
Social Area Analysis is an important and helpful tool in understanding a city's socio-spatial attributes. However, its applications in urban planning or other related practices are scarce. The first reason for this is that the linear correlation methods implemented in the past years oversimplified the complexities of the reality, and the results often bring a pattern irrelevant to the real world. Another reason is that visualization of spatial patterns requires such a vast amount of labor that sometimes it is impracticable. This paper aims to propose a non-linear correlation method using Self Organizing Map (SOM) techniques to investigate the social areas within a city and then to project the data on a GIS for visualization. We select Beppu City, Oita Prefecture as a testing area, because there are not many studies on social area analysis of Japanese cities, and after this, we intend to expand our study to larger cities of this country.