Studying Crime Incidents for Safety Analysis Project Abstract:
San Francisco State Government has revealed all the crime incidents of the past ten years of San Francisco. Broadly, the data contained information in the form of dates of the incidents, the type of the crime and the exact location of the incident reported. We wish to analyse the correlation in the various attributes and try discover some patterns which could help the concerned authorities
and the decision takers to formulate strategies to try improve the safety of the citizens. We would try to predict crime-vulnerable-areas dependent on the time using the available data. If successful, this could help anticipate the crimes and realise a dream of crime-free state.
The data consists of records of over hundred thousand reported incidents with each record consisting of attributes such category of crime, location, time, etc. Introduction In present world scenario, where major cities are prone to crime in great numbers, the crime analysts have to carry out a rigorous task of studying the crime statistics and assisting the authorities to cause arrests and crime prevention. However with the vast expanse of data and the lack of awareness about the locations of a city makes it even difficult for human to pursue the analysis tasks.
There may well be some inherent patterns in the reported incidents which get missed by ordinary human perception. There are various machine learning and data mining tools and techniques which study data and carry out fruitful results of classifications and prediction on a data well analogous to a crime database. The benefits of such an analysis can be huge, if we think about the full potential of the results and findings. Say, for example, if using the patterns one could predict that there could be sale of drugs in the southern district of the city at the end of this week, then this information can be utilized to highten the security of that area. Realizing the potential of a well maintained crime database, the police authorities around the world have taken steps to maintain quality data of the crimes reported.
San Francisco State Government sits on the top of such intelligent institutions who have openly published their data to the world to either use it for one’s personal statistical analysis or may help the state by notifying about certain patterns. The data from the past ten years is as voluminous as 100,000 records with as many as ten different attributes defining a record. Its not only inefficient but daunting task to carry out such an analysis by the police analysts. One requires a machine to recognise the patterns in the data accurately using a predefined metric. What we do here is no less than the prescribed needs.
Download Studying Crime Incidents for Safety Analysis Project Report