ABSTRACTActivity spaces are places in which people participate in an activity during a certain time period, serving as the starting or ending point of a trip. Previous studies on trip generation or attraction models have been limited to aggregated spatial and temporal scales. Extending recent research on the application of mobile phone data to transportation models and particularly addressing the above-mentioned limitations, this study investigates temporal and spatial characteristics of hourly activity space attractiveness by activity type at finer spatial (e.g. census block) and temporal (e.g. hourly) resolutions. We construct hourly activity space attraction models in the daytime for four activity types, including work, private education, shopping, and recreation in Gangnam, Korea. The models are log-transformed or square-root-transformed multiple regression models. We also test the spatial transferability of the Gangnam model to the area of Seocho, Korea.