Human mobility models reveal the underlying mechanism of seasonal movements across China

Understanding the spatial interactions of human mobility is crucial for urban planning, traffic engineering, as well prevention and control infectious diseases. Although many models have been developed to model mobility, it not clear whether such could also capture traveling mechanisms across different time periods (e.g. workdays, weekends or holidays). With one-year long nationwide location-based service (LBS) data in China, we investigate spatiotemporal characteristics population movements during periods, make thorough comparisons applicability five state-of-the-art models. We find that flows show significant periodicity strong inequality temporal distribution. A “backflow” effect found cross-city before after holidays. Parameter fitting gravity reveals travels type days consider attractiveness destinations cost distance differently. Surprisingly, comparison indicates parameter-free opportunity priority selection (OPS) outperforms other best characterize China all six types days. However, there still an urgent need development more dedicated on