Predicting Denver crime with linear regression : using tree canopy, poverty, & community demographics

Material Information

Predicting Denver crime with linear regression : using tree canopy, poverty, & community demographics
Montepagano, Mitchell
Younkes, Alexandra
Place of Publication:
Denver, CO
University of Colorado Denver
Publication Date:

Record Information

Source Institution:
Auraria Library
Holding Location:
Auraria Library
Rights Management:
Copyright [name of copyright holder or Creator or Publisher as appropriate]. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Auraria Membership

Auraria Library
Data to Policy Project Collection


This item is only available as the following downloads:

Full Text


Predicting Denver Crime with Linear Regression: Using Tree Canopy, Poverty, & Community Demographics This project investigates the relationship between crime rate in Denver and community demographics including: poverty, tree canopy coverage, number result of this investigation is a predictive model that estimates average crime model are useful explanatory variables for crime rate, but do not imply any city planners to ensure adequate resources are allocated to account for the continued research between crime rate and community demographics and may provide city planners insight into community support strategies that can Keywords: Linear Regression Analysis, Crime, Demographics Mitchell Montepagano & Alexandra Younkes University of Colorado, Denver Department of Statistical & Mathematical Sciences Abstract Tree Canopy The tree canopy dataset has over 385,231 observations of the was summed for all the trees in the neighborhood and converted to square neighborhood with its associated tree canopy in square miles and number of Crime The Denver crime data set contains 443,628 observations, each location (longitude and latitude), neighborhood, time, offense type, whether Foreclosures American Community Survey 2015 Data Methods Our research examined 70 variables , which translates to 2 70 or 1.2x10 21 possible models 1 trillion is 1 x10 9 combinations? Using the model selection methods involving selection 2 ), we built many models and then analyzed their predictive power to select the best model for the prediction of crime rate in The best models from these processes were compared against our best model found by backwards selection, or the process of systematically removing the Using cross-validation, the backwards selected model with 8 predictors was A unit increase in X is associated with a multiplicative change of e in the response, y: Coecients Estimate Change in Crime Rate Per Unit increase in X Intercept Tree Canopy (mi 2 ) # Liquor Licenses # Parks # People per Park % Poverty dents Connect Students And Experts: Data is insightful, providing clues on the relationship between community demographics and crime rate, but our analysis stops short of proving causal Expert analysis from community leaders, such as economists, sociologists, and criminal psychologists adds actionable power to Fund Robust Research: This research lies an excellent foundation for causal research between crime rate and community demographics, tree canopy coverage, number of them beautiful! Thus, we recommend that the city plant trees Our model indicates that there is a strong negative relationship between tree outdoors and provide residents with a sense of community and beauty, while also potentially decreasing crime rate, this could be a simple, and rewarding Another policy change we recommend, is alcohol education or access to alcohol, have community classes on recognizing alcohol dependence in loved ones, or educating the public on how exactly alcohol can affect your Lastly, poverty has been a known root of criminal activity since the dawn of opportunity can be the trigger rapidly changing infrastructure and increasing population, we tend to forget what kind of opportunities people need access to, but the idea that there is Policy Recommendations Works Cited