SimMiner: A Tool for Discovering User Similarity by Mining Geospatial Trajectories

This paper describes a tool called SimMiner that can model, analyze and visualize geospatial location information and discover similarity patterns from trajectory databases. The location data is modeled as a grid with cell transitions, which not only helps in mining and retrieving the similarity pattern between users, but also reduces the complexity for storage of such location information which is normally modeled as trajectories in moving object databases. The system also showcases an interactive browser-based visualization component that demonstrates the results of each step - modeling the grid, extraction of cell transition sequence and user similarity mining.