Demo Paper: A Spatio-Temporal-Textual Crime Search Engine
Xutong Liu, Changshu Jian, Chang‐Tien Lu · 2010
This paper proposes a STT(spatio-temporal-textual) search engine for extracting, indexing, querying and visualizing crime information. Until recently, it’s a labor-intensive work to identify crime entities, cluster similar suspect activities, and discover patterns from massive online collections. It’s a big challenge to reveal inherent ST(spatio-temporal) correlations among mass crime information. It’s getting more difficult considering the subjectivity and vagueness of information retrieval from narratives of victims or witness and online documents of social networks. We have developed a crime search engine for Washington DC metropolitan area that includes geo-temporal-tagger, STT indexer, heuristic query and ranker and dynamical ST visualization. It assists crime detection for investigators, identification of crime trends and patterns for decision makers and researchers, and security of city life for residents and journalists.