Detecting spatial patterns of incidents and crime trends is a topic of interest to any state or local government.
CAST is the name of a free software, initials of Crime Analytics for Space - Time, which was released in 2013 as an open source solution for actuarial analysis, with spatial patterns and algorithms of trends in the handling of crime statistics.
CAST is a client application developed on Python and C ++ that works on Windos, Mac and Linux, developed by no less than the GeoDA Center, which has developed various computational and spatial analysis applications. This Center has a laboratory that was founded by the Director of the School of Geographic Sciences and Urban Planning of the University of Illinois.
In CAST's case, it was promoted through an award from the National Institute of Justice and the Office of Justice Programs of the United States Department of Justice. The methodology for the development of algorithms was worked on with the Arizona State University.
The application supports SHP files, the incidents usually at the point level and through spatial analysis generates trends from dates, for which polygon maps such as neighborhoods, blocks or neighborhoods are required.
As results can be apart from the charts, thematic maps from statistical deviations, also heat maps and calendar maps.
Perhaps the most attractive thing about the application is that it comes with specialized functions already defined for the purposes of trend analysis and subject-based reporting. For example, a trend can be normalized by crossing the population data to represent the number of violent incidents in segments, as an example, the case of the number of deaths per hundred thousand inhabitants. Then it allows to do temporal analysis, to determine through graphs the growth, decrease and specific cases of study both at the tabular and spatial level. Similarly, personalization of the calendar can make analysis between specific days, such as incidents on holidays or weekends.
You have to play with the tool, because you can even generate animated maps on a time scale, with which it is possible to determine where a crime spot will spread if trends continue. Of course, it should be interesting to apply new data from security measures taken, to see the impact that is had. Something very useful in urban areas with the current context of the influence of organized crime and gangs that can surely be detected as interconnected clousters. And since the system is made for this purpose, it adapts to models of security management and violence prevention, such as the management of quadrants, sectors or districts.
In conclusion, a valuable application. One more under the open source model, to which we wish diffusion sponsors, without considering what governments invest in security for not so specialized functionalities.