Analyzing the Factors that Affect and Predict Employment Density Using Spatial Machine Learning: The Case Study of Seoul, South Korea

There is a regional disparity in the employment density of Seoul. Considering problems such as traffic congestion and jobs-housing imbalance, it is important to understand the spatial pattern of employment density and identify key influencing factors to determine the changes in the future urban spatial structure. This study analyzed employment density in each region of Seoul to derive important predictors. We examined the spatial patterns of employment density and evaluated the effects of spatial and nonspatial factors based on the general model and the spatial heterogeneity model. To predict the distribution of employment density, we used two statistical models (i.e., ordinary least squares regression [OLS] and geographically weighted regression [GWR] models) and two machine learning models (i.e., the random forest [RF] and geographically weighted random forest [GWRF] models). The results showed that the key influencing factors were the number of corporate business companies, number of main and attraction facilities, accessibility to subway stations, areas of commercial and industrial districts, and distance to business districts.