Discovering Homogeneous Geographic Areas Using Clustering

• Activity Number: 215
• Type: Contributed
• Date/Time: Monday, August 3, 2009 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Jin Yan Shao*+ and Bin Zhang and Ming Xie and Li Xia and Wenjun Yin and Jin Dong
• Companies: IBM China Research Laboratory and IBM China Research Laboratory and IBM China Research Laboratory and IBM China Research Laboratory and IBM and IBM China Research Laboratory
• Keywords: GIS ; data mining ; clustering ; homogenous areas

Abstract

The advance of Geographic Information System is providing more and more opportunities for geographic data analysis. In this paper, we deal with a practical application scenario: discovering the homogeneous geographic areas in a city based on GIS data. In many city GIS data analysis cases, it needs to study the sub-areas instead of just considering the whole city. For example, for locating the supermarkets in one city, it is necessary to divide the whole city into several sub-areas, and find the best locations for opening supermarkets in each sub-area, such as political districts. However, those "political-like areas" may not be suitable for the analysis work. Our method is to find out the homogeneous areas, each of which has specific characters. We use an aggregation way: firstly divide the city into small grids; secondly cluster the similar grids to generate the homogenous sub-areas.