Analyzing temporal mobility pattern shifts: Dynamic origin–destination estimation from aggregated mobile phone location data

• Spatio-temporal population statistics enable dynamic trip model estimation. • Gravity model with adjusted time-varying trip generation calibration steps. • Addresses limitation and potential of aggregated mobile phone presence data. • Day-to-day comparison contrasts with trip changes; intraday reveal hotspots. • Identified interpretative application on mobility pattern shifts through pandemic. Understanding the dynamic shifts in population distribution is crucial for comprehending spatiotemporal mobility patterns and phenomena at the urban scale. This is particularly relevant when examining shifts in the origin–destination trip patterns, as changes in the dynamics may signify the impact and directionality of mobility flows across areas. This study focuses on data-driven research utilizing population dynamics data, more widely known as aggregated mobile phone location data, to estimate the origin–destination trip matrices in a more dynamic and contextual form. Departing from traditional static approaches to estimate origin destination, this research develops an enhanced gravity model framework to estimate time-varying trip matrices, enabling a more contextual and relevant understanding of mobility flows across different areas. The study then analyzes shifts in trip patterns from the estimated dynamic origin destination patterns, focusing on the impact of COVID-19 pandemic-related policies on urban travel flows in Kyoto city, Japan. By expanding the utilization and understanding of the dataset and method, this research contributes to the development of more contextual uses of trip pattern analysis in the fields of urban and transportation studies.