POI Based Urban Commercial Centers Identification and Classification

Abstract

Recently, urban commercial centers’ layout attracts great attention in research fields of urban planning and economics. This paper proposes an approach, which can identify the boundary of urban commercial centers and classify these centers into a hierarchical classes by using open data source POIs. Two non-parametric algorithm: modified ASCDT and the k-means clustering algorithm are used to measure the agglomeration effect of commercial industries, detect industry centers and commercial centers, and classify industry centers and commercial centers. As a case study, this approach is applied to Harbin, China and find out 374 commercial centers in four levels. To some extent, this result reflects the reality of commercial centers distribution of Harbin. This result provide the information support of layout planning and location decision-making for planners and investors. The non-parametric adaptive clustering algorithm and accessible POIs data are used to make the approach applying to extensive region and more cities.