Identification of retail agglomeration places from NTT townpage database and analysis of their size, functional structure, and distribution: A case study of tokyo prefecture

Abstract

This study aims to identify retail agglomeration places in Tokyo prefecture from NTT Townpage Database (TPDB) by applying spatial clustering and to analyse their size, functional structure, and distribution using GIS. Eleven retail and service functions necessary in daily life are analysed, which consist of 164 types of shop and service facilities in TPDB. There were about 250,000 shops (including service facilities) in Tokyo prefecture as of March 2007. The distances between shops are measured using longitude and latitude data in TPDB. A program is constructed to delimit retail agglomeration places in terms of spatial contiguity and agglomeration volume. Shops are considered to be contiguous when the distance between them is less than spatial search distance of 50 m. Retail agglomeration places are then identified when their agglomeration volume is greater than or equal to the threshold of 20 shops. The number of retail agglomeration places identified by assuming these two parameters totals 1,143. Retail agglomeration places are classified into five hierarchical levels on the basis of size. The highest level consists of 10 shopping streets including the Ginza. The number of shopping streets at the high level is 13, at the upper level 46, at the middle level 269 (including 10 shop clusters, i.e., commercial buildings), and at the low level 805 (including 130 shop clusters). It is found that 125 types of shop are located restrictedly in one of hierarchical levels. Thirty-two types are located restrictedly in the highest level of retail agglomeration places. Therefore they are referred to as the highest order of retail function. On the other hand, 28 types of retail function are considered to be high order, 11 types upper order, nine types middle order, and 45 types low order. It is also found that 39 types of shop are not located restrictedly in one of hierarchical levels but are located in all hierarchical levels. They may be referred to as the footloose type and comprise 40% to 45% of the total at all levels of retail agglomeration places. Two types of shopping cluster are identified at the low level. One has nearly same composition ratio of the highest-order function as retail agglomeration places at the highest level. The other has a high composition ratio for the low-order function. As the result that the location patterns were analysed for retail agglomeration places in the 23 wards of Tokyo and Tama area, the places at the highest, high, and upper levels show a dispersed pattern and those at the low level show an agglomerated pattern. This study also tries to represent the supply geography in Toshima ward as a sample study area. There are 39 retail agglomeration places in Toshima ward. There is one shopping street at the high level, four at the upper level, 10 at the middle level, and 24 at the low level including six shop clusters.