We can take incidents of crime and mark their locations on a digital map. They can then be grouped according to specified criteria, such as how close they are to each other, to discover hot spots of crime, with each spot/group coded by color and shape of marker. 
We can divide an area by a regular grid, and color code the areas/cells by the density of crime in that area. The darker the color, the more serious the crime situation.
We can mark crime hotspots with ellipses to indicate their extent. We can then compare the hot spot patterns between different periods to study their evolution over time, between crime types to study their differences in behaviour. We can check, for example, whether the crime patterns between the May 1 and October 1 Golden Weeks are similar, or whether pick-pocketing and robbery happen at very different places and times. We can study the correlation between crime and weather, convenience stores, banks, night clubs, MTR stations, … Whether striking at crime in one place simply drives them to another, and where, …
Some police officers in charge of intelligence are highly interested in our work. Research and development can be fun and practical too.
4 comments:
Interesting work! I hope such statistical research do indeed lead to crime prevention, and a still safer city. Isn't it great that the police can join forces with academia in applied research?
Yeah, it is really fun working with the many officers. Imagine speaking to a room full of sergents, inspectors, and superintendents, about 20 of them. The neat thing is they really thought the work that we did was useful for them, and wanted to work with us.
厲害~
那些圖是 PDA 版本嗎?
Actually, no. It was screen-dumped from a stand-alone program running on a PC.
Post a Comment