Crime in Los Angeles Since 2020

Yi Chen Wu, Matthew Xue, Fred Wang, Tim Wang

Introduction

Los Angeles is the second most populous city in the United States, contributing about a quarter of California's GDP and approximately four percent of the USA's GDP. However, the downside of being such a large city is the higher incidence of crime. We find it intriguing to analyze LA crime data from 2020 to 2023 to gain insights into various topics such as crime trends, patterns, and demographics within the city. We deem this topic significant because it can inform crime prevention strategies and policy making.

Analyzing crime data facilitates the identification of trends and emerging patterns, aiding proactive crime prevention efforts. This may involve implementing targeted intervention programs like neighborhood watches, youth outreach initiatives, and community engagement activities. Moreover, crime data analysis provides policymakers with evidence-based insights to develop policies and programs aimed at reducing crime rates and addressing the root causes of criminal behavior. Such initiatives may encompass community policing, crime prevention, education, and social services.

We believe that through trend analysis, geospatial analysis, and examination of crime occurrences at different times of the day, we can gain a comprehensive understanding of crime incidents in LA.

Data Overview

Our dataset contains information about crime incidents occurring in Los Angeles. It includes relevant details for each crime, such as the date reported, date occurred, time occurred, area name, victim's age, sex, weapon used, and location. The data was sourced from Data.gov, a reliable and reputable platform, and is regularly updated by the LAPD. We utilized the Pandas library to load the data, which comprises 28 attributes, including a mix of geospatial, categorical, and continuous data. With 769,354 observations spanning from 2020 to 2023, our dataset provides ample data for generating the visualizations we aim to create. The dataset uses about 182 MB of memory.

Map Overview

We can visualize theft crimes geographically using an interactive map. The map below shows thefts in Los Angeles neighborhoods in 2020, for demonstration purposes. To see thefts across all years (2020-2023), please click here.

There are five layers that can be explored. The first layer shows overall counts of theft data over the time period in each location. The remaining four layers show crimes over each of the years 2020 to 2023. We can see that theft crime activity most often happens around Downtown Los Angeles. We can also observe that there are areas that have little to no thefts. This could be due to a lack of reporting in those areas, for instance.

Monthly Crime Type Trends

The frequencies of different kinds of crimes may have seasonal trends. Here we can use a line chart to visualize the monthly counts of the top ten most reported types of incidents. This helps surface clear trends over time across different types of incidents. To filter by incident type, use the dropdown menu below.

We can observe that certain kinds of incidents do indeed follow seasonal patterns. Burglaries, for example, appear to be more frequent during the summer months, and less frequent during the winter months. However, some incident types do not appear to follow seasonal trends. For instance, it appears that there was a significant spike in identitiy theft incidents starting in 2022 and into 2023, but the occurrence is not very cyclical.

Heatmap Matrix

We aimed to analyze the variations in crime rates over the course of a day for each month in our dataset. The analysis was depicted through a matrix where the y-axis represents different hours of the day, and the x-axis aggregates the crimes for each month across all years in the dataset. Our findings indicate that crime rates peak after 12:00 PM, with a significant amount of crime occurring later in the day. Conversely, the lowest incidence of crime was observed between 4:00 AM and 6:00 AM. We noted a modest increase in crime rates from 2020 to 2022. However, beginning in January 2023, there was a noticeable decrease in crime rates.

Crime Incidents Reported by Police Stations

We can use a horizontal bar chart to visualize the percentage of crime incidents reported from 2020 to 2023 by each individual police stations. Due to the large size of the original dataset, a representative sample representing 5% of the data was utilized for visualization. This sample was extracted using Python's random sampling method.
Each police station in Los Angeles is responsible for handling crimes that occur within its jurisdiction. Therefore, the percentage of reported incidents per station serves as a significant indicator of the level of crime present in that particular area.
The primary focus of this visualization is to provide insights into the distribution of crime incidents across different police stations. Two key observations emerge from the visual representation:

77th Street Community Police Station stands out with the highest proportion of reported incidents, surpassing 8%. Such a high proportion of incidents being reported in the jurisdiction of this police station could warrant further investigation.
Rampart Community Police Station reported the lowest percentage of incidents, with about 1% recorded. Topanga and West Valley Community Police Stations each had 2% or less share of the incidents reported. It may also be interesting to look into why this could be.

Overall, this visualization serves to highlight the varying levels of crime across different regions of Los Angeles, providing valuable insights for law enforcement and policymakers to address crime prevention strategies effectively.

Types of Incidents

We can visualize the kinds of locations at which incidents are occurring using a word cloud of the incident report descriptions. The size of each phrase is determined by the frequency that the phrase appears across all incident reports.

word cloud of terms used in incident report description

The word cloud suggests that many of the reported incidents occurred along streets, in or near residences, as well as in parking lots or near businesses, for example. This can immediately give us a good working idea of where incidents might be likely to occur.

Conclusions

Many incidents occur during noon to early afternoon. Crime incidents can happen anywhere, either in public or private places. Some police jurisdictions, such as 77th Street and Southwest, could be riskier than other areas due to higher frequencies of crime incidents. Top three crimes are stolen vehicle, battery, and burglary. There was a significant drop in identity theft after 2022. Crimes usually happen on the street or in single family dwellings. Some future findings and information we want to explore in the future include whether demographics are correlated to crime incidents in different neighborhoods, as well as delays between incidents occurring and being reported.