Extraction of Areas Encompassing Communities in Which People Desire to Live Using Residential-Preference Questionnaire Data

Abstract

Currently, the mainstream online search method used by many real estate portal sites is to show properties in the order of their distance from the nearest rail (including subway) station (i.e., the closest property first, etc.). However, this method does not necessarily enable a person to search for the place of residence that “fits” that person in a logical manner. For example, if it were known that certain community (town) “clusters” had similar characteristics, this would enable a search method whereby selection is made according to community characteristics. In the present study, a network of desirable communities was constructed from desirable residential-preference questionnaire data. Using a modularity optimization method, an attempt was made to extract area clusters having similar characteristics, with closely resembling spatial and psychological distances. As a result, the Tokyo metropolitan area could be divided into 75 such clusters. Each cluster was subsequently characterized using correspondence analysis.

Keywords

• complex network
• community detection
• market segmentation
• correspondence analysis