Mapping AI Deployment Readiness Across NYC Community Clinics
A dataset and interactive map evaluating broadband and electricity infrastructure across all 311 NYC ZIP codes to identify where on-premise clinical AI can be deployed today.
Overview
Deploying AI in clinical settings is not just a software problem. Before a model can run on-premise at a community health clinic, the underlying infrastructure must support it: reliable broadband for data access and model updates, and stable electricity for continuous operation.
This dataset addresses a question that precedes most AI deployment conversations: where can on-premise clinical AI actually run today?
Dataset
We assembled infrastructure data across all 311 NYC ZIP codes, combining:
- FCC broadband availability records (Form 477)
- EIA electricity reliability data
- NYC DOHMH community health center locations
The result is a per-ZIP-code profile of deployment readiness, published as an open dataset on HuggingFace and visualized in an interactive Gradio map.
Key Findings
- Broadband availability is near-universal across NYC ZIP codes at the provider level, but reported speeds vary significantly in historically underserved neighborhoods.
- Electricity reliability metrics show meaningful variation across boroughs, with implications for always-on inference workloads.
- 74 ZIP codes contain at least one FQHC or community health center, forming the core deployment target population.
Access
Dataset and interactive map are publicly available:
- HuggingFace Dataset: Layered-Labs/nyc-clinic-ai-infrastructure
- Interactive Map: Layered-Labs/nyc-clinic-ai-infra-map
- GitHub: Layered-Labs/nyc-clinic-ai-infrastructure