For Data Scientists & ML Engineers · Health Care Facilities
Health Care Facilities Data for Data Scientists
Health Care Facilities data for data scientists: 9 datasets on one shelf. Every one delivered as API, files, or warehouse rows.
financial time series api for backtesting · alternative data for quantitative research · where to get training data for health care facilities models
API, files, or your warehouse. Daily, weekly, or hourly.
Is there a usable time series for modeling here?
The deepest panel is HealthData.gov - COVID-19 Hospital Capacity by Facility (Historical Time Series): one row per hospital per collection week across roughly 6,000+ facilities and about 230 weeks, carrying 127 columns suffixed _avg, _sum and _coverage for bed counts, ICU occupancy, prior-day admissions and critical staffing shortages, distributed as CSV, JSON, GeoJSON, XML and OData through the Socrata API. Treat it as a finished archive - reporting ended with the May 3, 2024 update, and HHS made no statistical adjustment for non-reporting facilities, so missingness is documented signal rather than noise. For longer horizons, UDS files reach back to 1999 per awardee and AHRF editions archive to 1999-2000 (county CSV/SAS only from 2019-2020 onward). Only 181 of the 1,744 datasets Datadory catalogs are static snapshots (10.4%), so a frozen panel you can pin for reproducible runs is a feature here.
How should you assemble the stack?
Layer the COVID-19 capacity archive when utilization dynamics matter, pinned at one snapshot. Use AHA Fast Facts totals as the annual sanity check, and route request-form sources through procurement before ingestion code exists. Licensing splits cleanly into lanes: U.S. Government commercial delivery terms for HRSA and HealthData.gov, the NPPES Data Dissemination Notice for CMS, negotiated terms for AHD and Definitive. This page is the health care facilities slice of all data-scientists resources.
Straight answers
Is there a financial time series API for backtesting on health care facilities data?
No market tape exists, but two series behave like one. The HealthData.gov capacity file is a completed facility-week panel of ICU occupancy, admissions and staffing shortages ending May 3, 2024, served through the Socrata API as CSV, JSON or OData; UDS adds annual per-awardee financial and utilization tables reaching back to 1999.
Can health care facilities data work as alternative data for quantitative research?
Yes - it moves independently of market data. County-level AHRF provider-supply features, HPSA shortage designations, daily-refreshed HRSA facility locations and facility-week utilization history give you exogenous regressors for real-estate, staffing and payer models.
Where can I get training data for health care facilities models?
Add AHA Fast Facts totals for label sanity checks before touching any contract-gated directory.
Rows before rollout
Sample rows from any shelf entry — the field dictionary and coverage notes ride along. If the shelf misses what you need, say so; sourcing requests are half our job.
Talk to us