NYC Complaint Network Analysis
Network Analysis of Complaints from 2019 to 2021.
imgId: 15

Map shows hot spot clusters of grand larceny complaints using Louvain community detection. The first visualization shows hot spots in 2019 and then in 2020 there is a decrease in grand larceny, most likely from the pandemic. Jumping ahead to 2021 we see grand larceny hot spots start to increase as more places begin to open.
imgId: 16

Data model design used for network analysis project to import data into Neo4J and create nodes and relationships.
imgId: 17

Cypher code to execute link prediction machine learning and make predictions about where future offenses will occur based on previous complaints and nearby locations.
imgId: 23

Comparison tool developed to show the change in amount of complaints over time for precincts in NYC.
Description
This project analyzed data published by NYC Open Data, which contains historic complaint data. For the purposes of the analysis the data was filtered by more recent dates, ranging between 2019 and 2021. It only focused on more serious types of offenses, that being felonies. The complaints occurred in 77 precincts within the 5 boroughs of New York City. The complaints in the dataset represent 65 categories that are more general classifications of offenses that can further be broken down into 369 more specific classifications. I used this data to develop a tool to identify patterns of complaints happening in different areas of NYC and show the patterns geographically on a map. The tool is flexible enough to analyze different offenses over the course of different periods of time. Since New York City has such a large population of people and a large area for police to patrol, it is important to spot trends of offenses to better allocate police resources. Time, location and the type of offense are critical variables as they answer the questions, when, where and what is happening in NYC. Having these tools will give them an advantage in knowing where new patterns of offenses are emerging and come up with a plan of action to reduce the numbers.
The main goal is to use the tools and analysis results to make better decisions and develop actionable items to help reduce crime in a timely manner. Some examples of improvements that this project could lead to is knowing where to increase the police presence in different areas, better communication between NYPD and communities about emerging crime patterns and identifying where to allocate funds towards technology such as adding more security cameras.
An additional use case I was interested in studying is how crime has changed during the COVID-19 pandemic. This was also one of the factors for the chosen date range of the data being from 2019 to 2021. This represents three different periods: pre-pandemic (2019), pandemic (2020) and vaccine/return to normal period (2021). The objective is that by analyzing these time periods, conclusions can be derived about if certain types of offenses are higher or lower than normal which could be a result of different factors during the pandemic such as restaurants public places being closed, reduction of tourists and people being quarantined in their homes.
Skills Used
- Neo4J
- Python
- ArcGIS