New Jersey Opioid Overdose Spatial Analysis

Overview

As a Research Affiliate at the Edward J. Bloustein School of Planning and Public Policy (2020–2022), I conducted grant-funded spatial analysis with the New Jersey State Policy Lab to identify opioid overdose hotspots across the state. This research transformed administrative data from the New Jersey State Police into actionable geographic intelligence for public health response.

The project addressed a critical question: Where are overdoses concentrated, and what spatial patterns can inform intervention strategies?

Data Source

The analysis drew on a unique dataset: Narcan deployment records from the New Jersey State Police. Since 2014, all state troopers carry naloxone and are required to administer it when encountering suspected overdose victims. This policy created a comprehensive, geocoded record of overdose incidents across New Jersey.

Working with the Office of Drug Monitoring and Analysis, I accessed three years of deployment data (2019–2021) to examine spatial patterns at a statewide scale.

Methodology

Hexagonal Spatial Indexing

Rather than analyzing data by traditional boundaries like counties or municipalities, I implemented Uber’s H3 hexagonal spatial indexing system. This approach offers several advantages:

• Privacy preservation: Hexagons aggregate individual incidents without revealing exact locations
• Uniform analysis: Unlike irregularly shaped administrative boundaries, hexagons provide consistent spatial units
• Machine learning readiness: H3 indices integrate seamlessly with predictive modeling pipelines

The analysis divided New Jersey into 627 hexagons, each representing a consistent geographic area for comparison.

Figure 1: Total Narcan deployment counts by hexagon (2019–2021)

Cluster Detection

I applied Local Indicators of Spatial Association (LISA) statistics using GeoDa software to identify statistically significant clusters. LISA analysis reveals four patterns:

• Hot spots: Areas with high values surrounded by high values
• Cold spots: Areas with low values surrounded by low values
• Spatial outliers: High-low or low-high patterns indicating local anomalies

Figure 2: Hot spots (red) and cold spots (blue) identified through LISA analysis

Key Findings

The spatial analysis revealed clear geographic concentration of overdose incidents:

Urban Concentration: The Camden metropolitan region and northeastern urban corridor emerged as persistent hot spots across the study period. These areas showed statistically significant clustering of high overdose rates.

Rural Cold Spots: Northwestern and southern rural areas consistently showed lower-than-expected incident rates, forming cold spot clusters.

Temporal Stability: Hot spot locations remained largely consistent across the three-year period, suggesting structural rather than transient factors driving geographic concentration.

Figure 3: Statistically significant clusters (p < 0.05)

Policy Implications

This research demonstrated that opioid overdoses in New Jersey are not randomly distributed—they cluster in predictable geographic patterns. This finding has direct implications for resource allocation:

• Targeted intervention: Prevention and response resources can be concentrated in identified hot spots
• Built environment analysis: The hexagonal framework enables future research connecting overdose patterns to neighborhood characteristics
• Cross-agency coordination: Geographic intelligence supports collaboration between public health, law enforcement, and social services

Technical Approach

The analysis combined multiple tools and techniques:

• R: Data processing, H3 indexing, and statistical analysis
• GeoDa: LISA cluster detection and spatial autocorrelation
• QGIS: Cartographic visualization and map production
• H3 Library: Hexagonal binning and spatial aggregation

Connection to Current Work

This research established a foundation for my current government analytics work. The experience of transforming administrative data into geographic intelligence directly informs how I approach data infrastructure challenges in state government—whether tracking veteran homelessness across regional hubs or designing performance measurement frameworks.

The project also demonstrated the value of academic-government partnerships: rigorous methodology applied to real operational data produces insights that neither sector could achieve alone.

Acknowledgments

This research was conducted in collaboration with Lt. Jason Piotrowski and Lt. Adam Polhemus of the New Jersey State Police, Office of Drug Monitoring and Analysis. Their partnership made this analysis possible by providing access to administrative data while maintaining appropriate privacy protections.

The work was supported by the New Jersey State Policy Lab at the Edward J. Bloustein School of Planning and Public Policy, Rutgers University.

Publication

The full research summary is available on the Bloustein School website.