Theme-Oriented Visual Analysis of Crime with Big Data
LI Daicha · Geo-information Science · 2014
Information visual analysis is one of the key technologies in big data. The advent ofbig dataera promotes the development of visualization techniques and also brings changes to the traditional crime analysis.The crime visualization could offer assistance to crime analysis in practice. However, they are separated in application. The primary challenge that crime visualization faces is how to analyze data features' heterogeneity, scale,timeliness and complexity. This problem can be resolved by applying visual analysis, which allows users to explore data of different types and dimensions, and to obtain more valuable information with high correlation through interactions. Public security data, in thebig dataera, is characterized by multi-source and heterogeneity, and multi-dimension and long temporal series. Based on the characteristics of the data and criminal analysis theory, this article mainly focuses on the visual content, the representing method and the interactive design of visual crime analysis combined with geo-visualization and information visualization technologies, such as Wordle,Story line, parallel coordinate and scatter plot matrices, etc. A series of topic-oriented visual analyses were proposed in this study, including visual analyses based on spatio-temporal trajectory data of serial crime, real-time criminal data, spatio- temporal data of criminal process, criminal time- series statistical data, descriptive crime texts, criminal multidimensional attribute data, and crime-related statistical data. Supports from criminal cases investigation, trend prediction, hotspot analysis and references of visualizing studies from other fields were also offered and discussed in this article.