Abstract:
MobilityDB, an extension to PostgreSQL and PostGIS, is used for the management of spatiotemporal mobility data. With the continuous growth in mobility datasets, utilization of MobilityDB is need of time for storage management and extraction of useful information from uncertain moving data. Due to complex architecture and configuration of MobilityDB on Debian operating systems and non-availability on Windows, the study focuses on three areas related to MobilityDB in Pakistan. (a) Exploration of MobilityDB using Abstract Data Types (ADT) and identification of various use cases in Pakistan. (b) Data compression by constructing trajectories from raw GPS logs (c) Spatiotemporal Trajectory analysis to produce decision support information followed by animated visualization with respect to Cluster and Frequent patterns. Big mobility datasets of TPL received in SQL format were migrated to MobilityDB followed by environment building. Preprocessing of mobility data was performed with respect to trips generation and outlier elimination. Generation of trips is performed with the help of vehicle statuses using custom build SQL curser utility. Finally spatiotemporal trajectory analysis performed using MobilityDB and visualization of extracted information carried out using an open-source plugin named ‘Move’ in QGIS. To make this study self-contained and easy for readers, all areas were documented with configurational commands, source codes, spatiotemporal range and aggregate queries along with brief description. The results were concluded with the help of spatiotemporal queries followed by animated visualization of cluster and frequent mobility patterns. Results revealed that MobilityDB successfully implemented on big datasets reducing 99% data size thus increasing the database performance and optimization. The study also identified various applicability areas in Pakistan considering real-life use cases based upon TPL’s mobility data.
