RT2text: Making Trajectory Summarization More Accessible through Text Generation
Abstract
Trajectory summarization is vital for mobility analysis, extracting key movement patterns from large trajectory datasets. RT2text is a rule-based method that converts summarized trajectory into readable text. It uses regular expressions for data extraction, rule-based analysis, and Jinja2 templating for dynamic text generation, ensuring coherent descriptions of movement patterns. By incorporating Natural Language Generation (NLG), RT2text facilitates intuitive insights into mobility trends. We demonstrate RT2text’s effectiveness in turning structured summarized trajectory into meaningful narrative, enhancing data accessibility for analysis and decision-making. The code for this work is available at https://github.com/RepresentantativeMAT/RT2text.
References
Machado, V. L. et al. (2024). A survey on the computation of representative trajectories. GeoInformatica, pages 1–26.
Machado, V. L. et al. (2025). Towards data summarization of multi-aspect trajectories based on spatio-temporal segmentation. Journal of Information and Data Management, 16(1):38–51.
Mello, R. d. S. et al. (2019). MASTER: A multiple aspect view on trajectories. Trans. GIS, 23(4):805–822.
Pugliese, C. (2024). Unveiling urban and human mobility dynamics through semantic trajectory summarization. In 2024 25th IEEE MDM, pages 259–261. IEEE.
Rocchietti, G. et al. (2024). From geolocated images to urban region identification and description: a large language model approach. In Proceedings of the 32nd ACM SIGSPATIAL, pages 557–560.
