Analysis of street crime predictors in web open data
- PDF / 1,833,344 Bytes
- 25 Pages / 439.642 x 666.49 pts Page_size
- 86 Downloads / 179 Views
Analysis of street crime predictors in web open data Yihong Zhang1
· Panote Siriaraya2 · Yukiko Kawai2 · Adam Jatowt1
Received: 20 August 2019 / Revised: 1 November 2019 / Accepted: 1 November 2019 / © Springer Science+Business Media, LLC, part of Springer Nature 2019
Abstract Crime predictors have been sought after by governments and citizens alike for preventing or avoiding crimes. In this paper, we attempt to thoroughly analyze crime predictors from three Web open data sources: Google Street View (GSV), Twitter, and Foursquare, which provides visual, textual, and human behavioral data respectively. In contrast to existing works that attempt crime prediction at zip-code level or coarser granularity, we focus on street-level crime prediction. We transform data assigned to street-segments, and extract and determine strong predictors correlated with crime. Particularly, we are the first to discover visual clues on street outlooks that are predictive for crime. We focus on the city of San Francisco, and our extensive experiments show the effectiveness of predictors in a range of tests. We show that by analyzing and selecting strong predictors in Web open data, one could achieve significantly better crime prediction accuracy, comparing to traditional demographic databased prediction. Keywords Crime prediction · Web open data · Image and text analysis
1 Introduction Crime creates negative impacts on people’s lives, and therefore the ability to predict crime has been sought after by governments, business owners, and citizens alike. By predicting the likely number of crimes on a street, governments can design police patrolling more Yihong Zhang
[email protected] Panote Siriaraya [email protected] Yukiko Kawai [email protected] Adam Jatowt [email protected] 1
Department of Social Informatics, Graduate School of Informatics, Kyoto University, Kyoto, Japan
2
Division of Frontier Informatics, Kyoto Sangyo University, Kyoto, Japan
Journal of Intelligent Information Systems
effectively (Camacho-Collados and Liberatore 2015), business owners can choose better business locations, and citizens can plan safer travel routes (Kim et al. 2014). We follow the hypothesis of geographical crime analysis research that seek to understand the environmental and local community features related to crime (Taylor et al. 1985; Graif et al. 2014). For instance, the broken windows theory states that the relationship between crime and environment such as visible signs of disorder and lack of maintenance (hence “broken windows”) encourage further crime including serious ones (Wilson and Kelling 1982). Given the emergence of various geographical Web open data in recent years, geographical crime prediction using Web open data has attracted a number of research efforts (Gerber 2014; Wang et al. 2016; Zhao and Tang 2017; Yang et al. 2017). Essentially, these research aim to predict crime occurrence from open data, often open government data. Existing works mostly use zip-code level geographical unit or grids of square kilome
Data Loading...
