11:30 AM - 11:45 AM
[070] Analyzing the effects of Walkable Environments on Sales Performance in Commercial Area based on Deep Learning Technology with Google Street View Panorama Image
Keywords:Sales Performance, Commercial districts, Walkable environment, GSV Image, Deep Learning
The selection of commercial districts and locations is an important factor to consider when starting a business, and is affected by the built environment (Choi and Shin, 2001). In addition, the built environment is closely related to the amount of walking (Ewing and Cervero, 2010) and the demand in the restaurant industry is mainly driven by walking (Moudon et al., 2006). In previous studies, empirical studies shed light on the effect of the street walkable environment at the unit of neighborhood on store performance (Kim et al., 2015; Seong and Choi, 2017). However, it’s been insufficient for the previous literature to examine the correlation between the store performance in commercial area and walkable environment at microscopic level of street pedestrians. Therefore, this study examined the effect of the street environment at sights of pedestrians on retail sales in Seoul. It is measured by Semantic-segmentation based on Deep-learning technique with Google Street View (GSV) panorama images. Furthermore, this study use multiple regression analysis with public data from Seoul city to analyze various factors that can affect store performance of restaurant industry in small and large commercial district. As a result of the analysis, sales in small commercial districts correlated with greenery, and openness was significant in large districts. It can be seen that it is important to create a well-constructed green space in small commercial districts, and the development density of buildings is an important variable in large commercial districts to increase sales. These findings may help planners and policymakers better understanding how to increase commercial sales performance and revitalize local economy with the walkable environment.