Evaluating Spatio-temporal Data Models for Trajectories in PostGIS Databases

Abstract

Research in transport, ecology, health and other fields stands to profit from an improved understanding of movement. As movement data availability improves, the need for appropriate movement data analysis increases. However, the limited support for modelling moving objects in GIS hampers data exploration and analysis. This paper discusses trajectory data models and their implementation in the open-source spatial database system PostGIS. We quantify the difference in performance between PostGIS default trajectory support, dedicated trajectory data models, and commonly used point-based data models. To the best of our knowledge, this is the first paper to evaluate PostGIS default trajectory support and compare it to a proposed dedicated trajectory data type from the literature. Our experiments include computing trajectory duration and length, temporal and spatial filters, extracting positions at a certain time, and visualizing trajectories in desktop GIS. We also discuss the limitations of, and potential for, contextual trajectories and moving area object trajectories. Our results show that PostGIS functions for moving point object trajectories are fast, reduce query complexity, and provide good indexing integration, especially concerning multi-dimensional indices; the results also reveal that trajectory data models outperform commonly used point-based data models.

Open Geospatial Tools for Movement Data Exploration

“…Geographic information systems (GIS) are commonly used due to their strong spatial data handling functionality. However, lacking support for the temporal dimension in current GIS limits their potential for movement data analysis (Graser 2018).…”

“…2. Furthermore, PostGIS provides some built-in support for trajectory data handling which is covered in detail by Graser (2018). QGIS provides a wide range of rendering options and good rendering performance.…”

“…The trajectory support implemented in PostGIS bypasses some of the restrictions of Simple Features by storing time information in the measure value of LineStringM features (Graser 2018). This approach makes it possible to create a single LineStringM feature that represents a whole trajectory where every point along the trajectory retains its timestamp stored in the measure value.…”

MovingPandas: Efficient Structures for Movement Data in Python

“…The temporal support for trajectories in PostGIS (Graser, 2018) uses a variation of the Simple Features concept. It stores movement data in LineStringM objects, where the measure variable M contains the temporal information.…”

“…As with terminology and concepts, however, there is no consistent framework for the classification of trajectory functions. Recent work includes Dodge et al's (2008) On the technological side, we will have a closer look at implemented functions for movement data in the spatial database PostGIS (Graser 2018), the moving object database Hermes (Pelekis et al, 2015), and the R package trajectories (Pebesma et al, 2018;Moradi et al, 2018 preprint). Since it is clear that each implementation was created for a different research or application focus, it is challenging to devise a systematic and objective means of comparison.…”