How Close is Too Close? Exploring the Relationship Between Bike Theft and Police Facility Distance in Toronto

Toronto
Crime
Linear regression
This is the final project for the course SOCC70H3: Models of the Social World at the University of Toronto Scarborough in Winter 2024 supervised by Prof. Ethan Fosse.
Author

Zehui Yin

Published

April 10, 2024

Abstract:

Bike thefts pose a significant deterrent to urban cycling. This research delves into the influence of police presence on bike thefts and the geographical scope of its deterrent effect. Leveraging bike theft records from 2014 to 2023 in Toronto, alongside the most recent Canadian census data, I employ two-way fixed-effect Zero-inflated Negative Binomial Regression models to scrutinize the impact of network walking distance to the nearest police facility on the number of bike thefts within each census tract. To ensure robustness, I present a variety of model specifications, including OLS multiple regressions, Poisson regressions, and standard Negative Binomial Regression. The findings indicate that a police facility within a 500-metre radius notably curtails bike thefts, with statistical significance at the 0.05 level. Yet, this deterrent effect wanes with increasing distance and becomes inconsequential beyond 500 metres. Furthermore, at distances exceeding 2 kilometres, the presence of police facilities does not influence the number of bike thefts. The study infers that police facilities exert a localized effect on bike theft prevention in urban areas. Consequently, merely increasing the number of police facilities city-wide is not an efficacious strategy for reducing bike theft. Policymakers are thus encouraged to consider alternative methods or strategies for enhancing crime deterrence.

This project addresses the following two research questions:

  1. What causal influence does the presence of police facilities have on the frequency of bike thefts in Toronto?

  2. Upon observing a deterrent effect, to what extent does the proximity of police facilities causally mitigate bike theft incidents in Toronto?

You can find the code used in this project by clicking here.

If the embedded PDF is not displayed properly, please click here.