fk759 PUI2016 Extra Credit Project Proposal

PROBLEM DESCRIPTION

In recent years, cities across the world have resorted to automated traffic enforcement cameras in order to free up human law enforcement resources for other tasks. The City of Chicago has both a red light and speed camera program, which issues fines to owners whose vehicles are caught driving over the maximum speed limit or entering through a signalized intersection after the signal has prohibited them from doing so. Since these cameras use photographic enforcement, they provide live video feed all over the city, available to law enforcement officials at a moment's request. I am interested in whether or not the installation of these traffic cameras have led to a subsequent decrease in general levels of criminal activity.

The question is as follows:

Does the placement of red light or speed cameras reduce the rate of reported criminal activity within the various political wards of the City of Chicago?

DATA

I will use the datasets below as the basis for my analysis. For all three sets, there is a recorded address, date and latitude/longitude coordinate set for each violation. The Crimes dataset also has a unique case number for each recorded incident (unique ID), date-time stamp, police district, rough street block number the incident occured and type/severity of offense. They are all released directly by the City of Chicago on the city's online data portal.

Datasets

Crimes - 2001 to present

Red Light Camera Violations

Speed Camera Violations

ANALYSIS

I will use spatial autocorrelation to find the relationship between placement of automated traffic cameras and their effect on local crime throughout the City of Chicago's wards. Since the cameras were all placed and activated by July, 2014 I will compare the crime data for two periods: the 24 months preceding July, 2014 and the 24 months afterwards. Significant increases and decreases in crime should show up on a map as hotspots and coldspots respectively, and the geography used would be the city's fifty wards, creating a chloropleth map of the wards reflecting positive and negative changes in local crime rates. Lastly I will attempt to use Moran's I, which will determine if the expressed pattern is random, clustered or dispersed.

REFERENCES

How Spatial Autocorrelation (Global Moran's I) works

(Getis 2010)

Intro to Spatial Data Analysis in Python

DELIVERABLE

My deliverable will be a statistical conclusion on whether or not the placement of the cameras has any significant affect on the change in crime rates in Chicago police districts.

References

  1. Arthur Getis, J. K. Ord. The Analysis of Spatial Association by Use of Distance Statistics. Geographical Analysis 24, 189–206 Wiley-Blackwell, 2010. Link

[Someone else is editing this]

You are editing this file