Real-Time Social Media–Driven Crime Intelligence System Using Natural Language Processing and Geospatial Mapping

This paper presented the complete design, implementation, and empirical evaluation of a Real-Time Social Media–Driven Crime Intelligence System integrating NLP-based crime classification, spaCy Named Entity Recognition, Nominatim geocoding, and interactive Mapbox geospatial visualisation within a single deployable FastAPI/React prototype. The NLP engine achieved an overall F1-score of 85.9% across five crime categories. All fifteen functional test cases passed with 100% correctness. Qualitative usability evaluation identified five actionable design themes: interface clarity, intuitive interaction, perceived data value, conditional trust, and data freshness transparency. Comparative analysis confirmed that the proposed system is the only evaluated open-source prototype combining all pipeline stages—ingestion, NLP, geocoding, visualisation, and user evaluation—within a unified architecture. Future development directions include: (i) replacement of the keyword-regex classifier with a fine-tuned transformer model (RoBERTa or DeBERTa) trained on crime-specific annotated Reddit corpora to improve recall on vandalism and ambiguous categories; (ii) integration of confidence scores and keyword highlighting in the frontend popup to address the conditional trust theme identified in usability evaluation; (iii) addition of visible update timestamps, refresh indicators, and data freshness banners to resolve the data freshness transparency theme; (iv) multi-platform ingestion extending beyond Reddit to include X (formerly Twitter), local news RSS feeds, and police scanner APIs; (v) temporal filtering with a time-slider component enabling users to replay spatial crime patterns across configurable date ranges; and (vi) deployment on cloud infrastructure with horizontal scaling to handle sustained high-volume ingestion from multiple concurrent subreddit streams.

[1]  . Bernabeu-Bautista, L. Serrano-Estrada, V. R. Perez-Sanchez, and P. Marti, "The Geography of Social Media Data in Urban Areas: Representativeness and Complementarity," ISPRS Int. J. Geo-Information, vol. 10, no. 11, p. 747, Nov. 2021. [2] M. H. A. Abdullah, N. Aziz, S. J. Abdulkadir, H. S. A. Alhussian, and N. Talpur, "Systematic Literature Review of Information Extraction From Textual Data: Recent Methods, Applications, Trends, and Challenges," IEEE Access, vol. 11, pp. 10535-10562, 2023. [3]    D. Weisburd, "The Law of Crime Concentration and the Criminology of Place," Criminology, vol. 53, no. 2, pp. 133-157, 2015. [4]   H. N. Serere, B. Resch, C. R. Havas, and A. Petutschnig, "Extracting and Geocoding Locations in Social Media Posts: A Comparative Analysis," GI_Forum, vol. 1, pp. 167-173, 2021. [5]  B. Jehangir, S. Radhakrishnan, and R. Agarwal, "A Survey on Named Entity Recognition — Datasets, Tools, and Methodologies," Natural Language Processing J., vol. 3, p. 100017, 2023. [6]  R. Ahmad et al., "CHART: Intelligent Crime Hotspot Detection and Real-Time Tracking Using Machine Learning," Computers, Materials & Continua, vol. 0, no. 0, pp. 1-10, 2024. [7]  Md. M. Mukto et al., "Design of a Real-Time Crime Monitoring System Using Deep Learning Techniques," Intelligent Systems with Applications, vol. 21, p. 200311, 2024. [8]   M. S. Gerber, "Predicting Crime Using Twitter and Kernel Density Estimation," Decision Support Systems, vol. 61, pp. 115-125, 2014. [9]  G. A. Taiwo, M. Saraee, and J. Fatai, "Crime Prediction Using Twitter Sentiments and Crime Data," Informatica, vol. 48, no. 6, 2024. [10]  S. Ouni, F. Fkih, and M. N. Omri, "BERT- and CNN-Based TOBEAT Approach for Unwelcome Tweets Detection," Social Network Analysis and Mining, vol. 12, no. 1, 2022.