Welcome to My Final Prodject Page

This is a website built entirely with GitHub Pages and Markdown with HTML snipits for the visualizations Created by Drew Flinn and Iker Sanchez!

Introduction

Data and Data source

The dataset analyzed in this project is the “Chicago Crime Incidents 2001 to Present”, sourced by Harshdeep Sharma on the website Kaggle, extracted from the Chicago Police Department’s CLEAR (Citizen Law Enforcement Analysis and Reporting) system. This data represents a public record of reported incidents of crime (excluding murders, where data exists for each victim) that occurred in the city of Chicago. The dataset is quite extensive, comprising approximately 8 million records with very few instances of missing data. Each record represents a distinct incident and includes key variables, such as the date and time of the occurrence, block-level address, crime type (following IUCR codes), location description (street, residence), arrest status, and domestic classification. Crucially for our spatial analysis, the data includes latitude and longitude coordinates for the majority of incidents.

Research Questions

This analysis attempts to answer two questions regarding the evolution of crime in Chicago:

• How have the geographic locations (“Hotspots”) of crimes changed from year to year?
• Has there been a significant change in the overall volume and specific types of crimes (“Trends”) committed over the years?

With the first question, we aim to observe if crime remains concentrated in specific neighborhoods or if it fluctuates due to other confounding variables. With the second question, we aim to identify trends in the most frequently reported crimes and observe how their prevalence has changed over the two decades.

Analysis Approach

Our approach involves a combination of time-series aggregation and geospatial visualization, for our first research question we used the geospatial visualizations and for the second we focused on time series aggregation and analysis.

Question 1

Link to Google Colab

For this first question we used a scatter mapbox in conjunction with binning To create the visualization.

Data Preparation Steps:

• Drop rows without coordinates.
• Converting each crime’s location into an H3 hexagon.
• Group incidents by year and by H3 cell.

We initially attempted a standard heatmap using all ~8 million inputs, However we repeatedly ran into memory issues with colab and due to the size of the data every time we had to restart it added about 3 minutes just for it to iterate over the data. Switching to H3 binning enabled us to go through section by section and find like points instead of plotting each of the 8 million points.

Question 2

Link to Google Colab

For this question, we first did some data management to extract the wanted information and group it depending on what we wanted.

Visualization 1

Visualization 2

###Data Preparation Steps:

• Data set was grouped by “Year”.
• Calculated the total count of data points per year, so the total number of crimes per year.
• Results were sorted by index to ensure chronological order.
• Value counts function was used on “Primary Type” to identify the most frequent crimes.
• Top 5 crime types were filtered for further analysis.
• A new df showing the annual trends was generated by grouping by “Year” and “Primary Type”.

We initially ran into an issue in which the numbers we were obtaining yearly were making little to no sense as they were not even reaching the thousands, and we knew that the dataset contains around 8.5 million data points so clearly there was something going wrong. We realized the issue was that the data set was not being loaded properly, so we had to change the approach to how we were introducing the data in colab. After figuring that out we had no other issues.

Results:

The animated visualization reveals that Chicago’s crime hotspots show geographic stability across the 2001–2025 period. The brightest and most intense hexagons appear in the central, western, and southern areas of the city. These areas display the highest crime counts each year in the animation.

Individual hexagons change color from one year to the next, reflecting changes in annual crime totals however the overall geographic pattern does not shift. No new neighborhoods emerge as hotspots, and none of the existing hotspots disappear or move to different parts of the map.

The analysis of crime volume over the years reveals a generally decreasing trend, indicated by the negative slope in the visualization. However, specific periods showed notable deviations. From 2015 to 2016 there was a significant increase in crime. Potential contributing factors may include a decline in proactive policing (ACLU effect), increased gang activity, and socioeconomic challenges for minority groups. And from 2021 to 2023, years following the COVID-19 pandemic, we observe a spike in crime. This increase was expected as societal activities returned to normal and the population resumed their pre-pandemic routines.

In terms of the most reported crimes, we have Theft, Battery, Criminal Damage, Assault, and Narcotics (From most to least reported in 2025). The temporal analysis shows that while all top crime types followed the general decreasing trend, narcotics-related incidents saw a particularly sharp decline in the last 2 decades. This may reflect shifts in law enforcement priorities regarding drug offenses.

Conclusion:

Based on the results of the visualization, crime hotspots in Chicago do not move locations over time. From 2001 through 2025, the highest-crime areas remain in the same areas of the city, with no evidence of movement or the creation of new high-crime zones. The only changes observed are fluctuations of crime levels within the already established zones. Beyond this spatial consistency, the temporal analysis reveals a significant overall decline in reported crime volumes from 2001 to 2025. Despite this positive long-term trajectory, specific challenges persist, particularly with the sustained high frequency of Theft and Battery incidents. On the other hand, Narcotics offenses have seen a steep reduction over the last 2 decades, potentially reflecting shifts in enforcement strategies or underlying activity. Ultimately, while the locations of crime in Chicago have remained static, the composition and volume of these incidents have evolved, underscoring the need for policy interventions that address these changing dynamics within the city’s established hotspots.