9:40 AM - 10:00 AM
[D1-03] Applicability of alternative data for the identification of urban activity centers (UAC)
Keywords:urban activity, urban centers, spatial modelling, machine learning, alternative data
The spatial distribution of urban activities and places of their concentration in cities (urban activity centers – UAC) is one of the main issues within Urban Spatial Studies. However, due to the inaccessibility of corresponding data, the data not related to urban activities directly were used to detect and analyze UAC in most previous works. In this research, we use the Person Trip Survey results for 23 special wards of Tokyo as the data describing urban activities directly to clarify the previously understudied relationship between the actual activities and alternative data commonly used to analyze them. Given the availability and relevance to the study area, Twitter data, Building Point data by Zenrin and POI of OpenStreetMap are used. This study clarifies the applicability of urban activity indicators extracted from alternative data by means of UAC modelling with machine learning algorithms. Its findings allow us to evaluate the accuracy of previous studies where the relationship between the targeted phenomenon and the data used is not illustrated and to provide guidance on data selection for urban activity and UAC studies.