11.522: UIS Research Seminar (Fall 2014) - Discussion notes
After 2007, Mexico underwent a severe violence crisis related to drug cartels. From 2007 to 2011 rate of homicides went from 9 per 100 thousand population to 24 per 100 thousand inhabitants. Neither the security institutions nor the information technologies of the country were prepared to tackle this challenge (See Note). In particular, crime databases both at the local and national level (including homicides) were poorly maintained and managed. This was a result not only of a lack of human capital, but also of the political sensitivity of the data: violence and crime became the focus of the country.
Despite this lack of crime data in Mexico, it was clear from the beginning that the spatial distribution of crime was not random, but instead was clearly distributed at the state and municipality level. (Mexico has three government levels: 1 federal (national), 32 states and over 2,400 municipalities.) However, systematic analysis of the spatial dimension of crime has remained relatively underdeveloped. This is all the more troubling given that the administration of President Enrique Pena Nieto (2012-2018) launched a National Program for the Social Prevention of Violence and Crime in which it selected municipalities to channel resources designated for urban and social development initiatives. The methodology for this municipality selection was not made publicly available and critics pointed out that they did not correspond to the localities in more need of these resources (given their social vulnerability and/or crime rates).
As part of my research, I would like to develop a basic spatial analysis using GeoDa software in order to identify clustering patterns of crimes in recent years in Mexico. I am particularly interested in kidnapings at the municipality level, as the data might be available soon and this has become the highest growing crime in 2014. I would like to study kidnapings in relation to other social cohesion and economic data available also at the municipality level. My hypothesis is that higher levels of social vulnerability are positively associated with higher crime rates, especially kidnapings.
For this class, however, rather than focusing on the data for Mexico itself, I want to think a little more broadly about the role of urban information systems in crime prevention and prediction. I am particularly interested in thinking about how cities should think about the collection of crime data and how to better leverage it: both for their own security tasks and for raising awareness and creating value in partnership with citizens and social organizations.
The lead author of the first reading, Luc Anselin, is the principal author of the GeoDa software that models spatial relationships and clustering. Skim this book chapter to get a sense of the methods and issues involved in doing a spatial analysis of crime. The second and third readings are journal articles. The second reading (by Pattavina, et al.) uses Boston as an example for discussing urban neighborhood information systems that could facilitate crime prevention and control. The third reading (by Sampson, et al.) is an article in Science that estimates and interprets a particular model of the relationship between violent crime and neighborhood characteristics. Be sure to read it in sufficient detail to understand the model, the evidence, and their interpretation.
Other readings and resources (mainly about Mexico’s National Program for the Social Prevention of Violence and Crime):
Back to 11.522 home page.