Location Analysis of Retail Stores in Changchun
Total Page:16
File Type:pdf, Size:1020Kb
Cities 41 (2014) 54–63 Contents lists available at ScienceDirect Cities journal homepage: www.elsevier.com/locate/cities Location analysis of retail stores in Changchun, China: A street centrality perspective ⇑ Fahui Wang a,b, , Chen Chen c, Chunliang Xiu c, Pingyu Zhang b a Department of Geography & Anthropology, Louisiana State University, Baton Rouge, LA 70803, USA b Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, Jilin 130102, China c School of Geographical Science, Northeast Normal University, Changchun, Jilin 130024, China article info abstract Article history: This paper examines the location pattern of various retail stores in Changchun, China. The centrographic Received 20 December 2013 method, the nearest neighbor index and the proximity to CBD are used to provide some baseline analyses Received in revised form 14 May 2014 of their spatial distributions. Major findings are derived from the street centrality indices measured in Accepted 18 May 2014 terms of a node’s closeness, betweenness and straightness on the road network. The kernel density Available online 17 June 2014 estimation (KDE) converts both store locations and centrality values at nodes to one unit (raster pixel) for correlation analysis. Results indicate that street centrality captures location advantage in a city and Keywords: plays a crucial role in shaping the intraurban variation of commercial land use intensity. Specifically, Street centrality specialty stores value various centralities most, followed by department stores, supermarkets, consumer Retail stores Location preference product stores, furniture stores, and construction material stores. Among the stores with correlation Kernel density estimation coefficients above 0.5, specialty stores favor closeness most, department stores and supermarkets prefer Correlation betweenness, and consumer product stores value straightness most. Ó 2014 Elsevier Ltd. All rights reserved. Introduction From a vendor’s perspective, ‘‘being central’’ is a major principle that can be traced back to the Hotelling’s (1929) classic location ‘‘No matter how good its offering, merchandising, or customer problem of ice-cream shops. The intuitive notion of central ten- service, every retail company still has to contend with three critical dency is further advanced by the space syntax analysis (Hillier & elements of success: location, location, and location’’ (Taneja, 1999, Hanson, 1984) and the complex network science (Barabási, 2002; p. 136). Location analysis is a common and important task in busi- Batty, 2008), specifically various centrality indices based on a ness management (e.g., Ghosh & McLafferty, 1987; Berman & street network. As Hillier, Penn, Hanson, et al. (1993, p.32) state, Evans, 2001; Zentes, Morschett, & Schramm-Klein, 2011), as exem- non-residential economic and service activities in urban neighbor- plified by the profound influence of the Huff (1963; 2003) model in hoods have been found to be ‘‘determined by the structure of the market studies. urban grid itself rather than by the presence of specific attractors While the importance of location for retail stores is no question, or magnets.’’ In other words, it is the configuration of a city’s street the assessment of location is not an easy task. The advancement of network that shapes its economic or social dynamics and structure. Geographic Information Systems (GIS) has helped develop and A primary focus of the space syntax approach is centrality operationalize some quantitative measures of location. This measures derived from the street network and examination of includes various indices of accessibility from the earlier potential their association with various economic activities. model by Hansen (1959) to the gravity-based availability measure There is a rich body of literature examining how urban land use (Weibull, 1976; Joseph & Phillips, 1984), and to the more recent or economic activity patterns are closely associated with various two-step floating catchment area method (2SFCA) and its general- centrality indices. The approach defines a place ‘‘being central’’ ized form (Luo & Wang, 2003; Wang 2012). The accessibility not only in terms of closeness (proximity) to other places as in tra- approach is mainly from a consumer’s perspective in order to ditional geography, but also being ‘‘intermediary, straight... and capture the convenience of a resident reaching or obtaining a critical’’ to others (Porta, Crucitti, & Latora, 2006). Therefore, it is service offered at various facilities or ‘‘attractors’’. a more comprehensive assessment of location. Most recently, Porta et al. (2009) implemented several network-based centrality ⇑ Corresponding author. Tel.: +1 225 578 6629; fax: +1 225 578 4420. indices, termed as the multiple centrality assessment (MCA) E-mail address: [email protected] (F. Wang). model, to capture location advantage of various places and found http://dx.doi.org/10.1016/j.cities.2014.05.005 0264-2751/Ó 2014 Elsevier Ltd. All rights reserved. F. Wang et al. / Cities 41 (2014) 54–63 55 them highly correlated with distributions of commercial and ser- of Commerce (in collaboration with the Changchun Institute of vice activities in an Italian city. The same set of indices is used to Urban Planning) (2011). We expanded the list by using the Baidu explain variation of land use intensity such as population and Map (map.baidu.com), a popular search engine in China similar employment densities (Wang, Antipova, & Porta, 2011). Porta, to the Google Map, to search for stores in Changchun by various Latora, Wang, et al. (2012) analyzed the association of centrality categories. The final data set includes 973 retail stores in the study with a wide range of economic activities and found that the corre- area. Each record contains a store’s name, address and business lations are higher with secondary (e.g., services) than primary (e.g., type. We are aware that the stores obtained by this approach are manufacturing) activities. not an exhaustive list of all retail stores in the study area. The However, according to our knowledge, no prior studies exam- Baidu database only maintains significant stores based on the ined whether various types of retail stores tend to be associated number of times a store being searched for, and thus includes usu- with one centrality measure more than others. In other words, ally large and notable stores. Many more stores are small with a does the location preference differ among various categories of short lifetime and not included in this study. Nevertheless, it is retail stores? Furthermore, there are few case studies from devel- the best approach that is feasible for the research team. There oping countries that apply the analysis of centrality in an intraur- are many classification schemas for retail stores and commercial ban setting. Most studies examine the relationship between outlets (Guy, 1998). This study follows the conventional Chinese centrality of transportation networks and regional development government classification (also adopted by Changchun Bureau of patterns such as in China (e.g., Li & Cai, 2004; Wang, Jin, Mo, & Commerce) with six major categories: specialty stores, department Wang, 2009; Wang, Mo, Wang, & Jin, 2011) and India (Bagler, stores, supermarkets, consumer product stores, furniture stores, and 2008). One exception is the study by Gao, Wang, Gao, and Liu construction material stores. Specialty stores are those specializing (2013) that used street betweenness to explain traffic flows in in a specific range of merchandise and related items such as clothes Qingdao, China, but its focus was not on locations of economic and footwear, electronics, books, pharmacies, spectacles, stationer- activities. There are also some intraurban studies on the location ies, toys, etc. Consumer product stores sell products and commod- patterns of various firms in China, but the case studies are limited ities in retail or wholesale for common household needs such as in scopes such as on high-tech industries (Zhang, Huang, Sun, & produces, meat, seafood, other groceries and general merchants. Wang, 2013), industrial firms in general (Qi, Fang, & Song, 2008) Construction material stores contain steel, pebble, gravel, sand, or retail stores in downtown area (Chai, Shen, & Long, 2007). None concrete, mechanical equipment, and other construction supplies. examined the association with centrality indices. The other three types are self-explanatory. This case study examines the location pattern of various retail The street network is based on a base map used by the stores in Changchun, China. According to the Changchun Bureau Changchun Bureau of Commerce (2011). The locations for retail of Commerce (in collaboration with the Changchun Institute of stores are obtained on the Baidu Map by searching for their Urban Planning (2011), retail stores are classified into six catego- addresses in Chinese one by one, and then input into ArcGIS. Like ries, namely specialty stores, department stores, supermarkets, many cities in China, Changchun does not have a typical central consumer product stores, furniture stores, and construction mate- business district (CBD) as in most Western cities despite a series rial stores. Our emphasis is whether the six types of stores display of public campaigns for planning and building one (http://finan- different location preference from the perspective of street central- ce.ifeng.com/money/roll/20090723/978630.shtml). The place that ity. By doing so, we are not only interested