Higher Prevalence of Surveillance Cameras in Racially Diverse Neighborhoods across 10 US Cities

This repository contains code and scripts used in the paper "Higher Prevalence of Surveillance Cameras in Racially Diverse Neighborhoods across 10 US Cities". The study leverages computer vision techniques to analyze surveillance camera prevalence across various neighborhoods in 10 U.S. cities, focusing on understanding the relationship between camera prevalence and neighborhood racial diversity.

---

System Requirements

Software Dependencies

This analysis requires R and the following packages (specific versions tested on indicated):

• tidyverse: 1.3.2
• pscl: 1.5.5
• conflicted: 1.2.0
• marginaleffects: 0.17.0
• ggpubr: 0.6.0
• sf: 1.0.12
• glue: 1.6.2
• tidycensus: 1.2.3
• broom: 1.0.4
• splines: 4.2.2
• scales: 1.2.1

Operating System Compatibility

The repository has been tested on:

• macOS Monterey

---

Installation Guide

Instructions

1. Ensure that R is installed on your system (version 4.2.2 or higher recommended).
2. Install the required R packages by running the following command in your R console:

   install.packages(c("tidyverse", "pscl", "conflicted", "marginaleffects", 
                      "ggpubr", "sf", "glue", "tidycensus", "broom", "splines", "scales"))

3. Clone this repository to your local machine.

Repository Structure

This repository contains three R scripts used to generate the figures in the paper. Below is a description of each script:

• descriptive-figures.R: Generates the descriptive statistics and visualizations, specifically Figures 1 and 5 in the paper.
• cs-results.R: Produces the results and figures highlighting the relationship between neighborhood racial composition and surveillance camera prevalence.
• change-results.R: Produces the results and figures for findings related to changes in surveillance camera prevalence.

Data

The datasets and codebook used in this study are publicly available in the Stanford Digital Repository. The data can be accessed at the following link:

https://doi.org/10.25740/jr882ny4955

The repository includes:

• Datasets: Camera location data, neighborhood demographic data, and related variables.
• Codebook: Detailed descriptions of the variables and data sources used in the study.

Camera Detection Code

All code for camera detection is here.

How to Run

1. Clone this repository:
2. Download the datasets from the Stanford Digital Repository.
3. Place the downloaded files in the data/ directory of the cloned repository.
4. Run the provided R scripts.

All figures will be saved in the outputs/ directory.

Expected Runtime

Each analysis script takes approximately 2-5 minutes to execute.