Directing resources where they matter most.
A data-driven view of overdose risk, built in partnership with university and public-health collaborators.
Research Intelligence Center
Exploring how data, public health research, and predictive modeling can support earlier overdose prevention and faster community response.
Simulated demonstration data only — not a real operational prediction system
How a Predictive Model Could Work
Public Health Data
Aggregated and de-identified information may include overdose calls, naloxone distribution, EMS response patterns, and broader community-level indicators.
Responsible Research Principles
- Use aggregated and de-identified data
- Never target or label specific individuals
- Avoid stigmatizing neighborhoods
- Test for algorithmic bias
- Include community partners in decision-making
- Validate predictions before real-world use
- Clearly communicate uncertainty
- Require appropriate legal, ethical, and institutional approval
Predictive tools should support trained public health professionals, not replace clinical judgment, emergency response systems, or community expertise.
Research Projects (Concept Development)
Overdose-hotspot prediction & public-health data
Atlas is developing a data-driven project to identify overdose hotspots — helping direct resources, education, and outreach to the communities where they are needed most. This work is built in partnership with university and public-health collaborators, and is under active development.
- 01Mapping overdose trends to guide resource allocation
- 02Built in collaboration with university and public-health partners
- 03Designed to inform, not replace, community-level response
Partner on the data that guides our response.
If your university or public-health department wants to collaborate on this work, we want to hear from you.