Abstract: Objective The spatial analysis of dynamic data faces uncertainty problems.This paper utilizes mobilephone positioning data as the example to identify the city center,and to explore the modifiable areal u-nit problem and the uncertain geographic context problem.We find that the spatial autocorrelation ofactivity density distribution is influenced by the zoning scheme and areal unit scale,while complexspatial and temporal dynamics also generate uncertainty.We also discusse possible means to mitigatethe uncertainty problems.