🏥 NYC Clinic AI Infrastructure Explorer

Research dataset for evaluating on-premise / local AI deployment at clinics across all 311 NYC ZIP codes.

Combines FCC broadband coverage (Jun 2025), US Census internet access data (ACS 2022), and EIA electricity cost & reliability data (2024) into a single ZIP-level dataset.

📂 Dataset on Hugging Face  ·  💻 Source code  ·  📄 CC BY 4.0
Explore metric

About This Dataset

Running AI inference locally on clinic hardware (rather than calling cloud APIs) can:

  • Keep patient data on-premise (PHI compliance)
  • Work in low-connectivity environments
  • Cost a fraction of cloud inference at scale

This dataset maps the infrastructure prerequisites (internet + electricity) by ZIP code so clinics and researchers can evaluate feasibility at a neighborhood level.

Key finding: NYC's electricity grid is excellent for 24/7 inference everywhere (Con Edison SAIDI: 14.9 min/yr). The real barrier is broadband access equity. South Bronx ZIPs with 55-63% fixed broadband are exactly the communities with the greatest need for reliable, local AI tools.