CMS provider data pipeline
Python ETL that loads nursing home provider data from data.cms.gov into SQL Server. Each run checks the latest processing date and loads only new data.
- Python
- SQL Server
- REST API
Data engineering · Analytics · Finance
I build the pipelines that bring data in, the reports teams run every day, and the financial analysis built on them. Since 2022 I've done that for a network of about 50 skilled nursing and assisted living facilities. Before that I worked in treasury, so I know what finance needs from the numbers.
Impact
Hours per cycle for finance processes I rebuilt with SQL, Python, and SSRS.
| Process | Before | After | Saved |
|---|---|---|---|
| Every month | |||
| Health insurance invoicingAbout 50 facilities | 32 hrs | 2 hrs | 30 hrs |
| Accounts payable run100+ vendor cash files | 8 hrs | 1 hr | 7 hrs |
| Monthly total | 40 hrs | 3 hrs | 37 hrs |
| Each cost-report cycle | |||
| RHCF cost report prep15 reports per facility, 50 facilities | ~1 week | < 2 hrs | ~1 week |
Tens of thousands
of dollars a year in incorrect billing stopped by compliance checks that flag inactive employees and duplicate facility charges.
$100K+
in owed funds recovered as a Treasury Analyst by working through HUD loan agreements, required facility improvements, and reimbursement approvals.
Selected work
Grouped by the kind of work. Most of it was built at Centers Health Care for finance, quality, and operations teams.
Pipelines, integrations, and automation that move data from source systems into clean tables and finished files.
Python ETL that loads nursing home provider data from data.cms.gov into SQL Server. Each run checks the latest processing date and loads only new data.
Python and SQL pipeline that produces 100+ vendor-specific cash management files each month and removes most manual data entry.
Matches provider files to internal SQL data through crosswalk tables, then generates SSRS invoices for about 50 facilities. Compliance checks flag inactive employees and duplicate facility charges before invoices go out.
Python script on the Jotform API that gives 9 adult day care centers a real-time census, replacing daily Excel tracking and monthly recaps, with automated reminder emails.
Also: a PBJ staff-turnover workflow with an exclusion table, prefilled review links, approval routing, and email reminders, and an insurance eligibility check that tests hours across three rolling weeks.
Reports, audits, and analysis that answer a specific question for the people making the decision.
Automated monthly CMS Star Ratings reporting with comparisons against 1, 12, and 24 months earlier. Tracks measure changes such as the planned FY2028 removal of the healthcare personnel and resident vaccination measures.
Recalculated health inspection scores for New York facilities using CMS methodology and found three calculation errors: Cycle 2 weighting, IDR-flagged citations, and an excluded F-tag. Also replicated Special Focus Facility scoring and the Five-Star staffing domain.
Multi-facility Medicare SNF length-of-stay report with rolling three-month benchmarks and readmissions built into the stay logic.
Analyzed 496,632 NYPD-reported collisions in Brooklyn from 2012 to 2026 and turned them into recommendations for a city council audience: leading causes, crash timing, the worst ZIP codes and streets, and vehicle mix.
Also: Power BI reports built with bookmarks and field parameters, weekly overtime reporting in SSRS, a PCC lab-orders dashboard, and antipsychotic-rate and discharge-to-home reporting.
Valuation, treasury, and cost reporting where the numbers have to tie out.
Python tool that turns a ticker into an 8-sheet Excel DCF built from SEC filings, with every forecast cell a live formula. Run end to end on Apple, Costco, P&G, and Ensign.
At Plainview Healthcare Partners: weekly cash-flow analysis for the CFO across 18 nursing homes, 30+ wire and ACH transfers a day, and $100K+ in owed funds recovered through HUD loan agreement review.
Fifteen SQL queries formatted in SSRS for Excel upload, replacing 15 hand-built reports per facility across 50 facilities.
Matching the operators behind Omega, Sabra, and CareTrust to CMS quality data to test whether star ratings and staffing show up in the rent coverage those REITs report.
Also: a cost audit of more than 450 shared mailboxes for the CFO, using a Jotform delegate review.
Side project · Live
A free web app for keeping a daily Torah learning schedule. Pick a program, mark each day learned or missed, see exactly where to pick back up, and count chazarah (review).
I designed and launched it on my own domain, hosted on Cloudflare Pages.
Open Keviut.com ↗Experience
2022 – Present
Centers Health Care · New York, NY · Promoted Feb 2025, Data Analyst Feb 2022 – Jan 2025
2021 – 2022
Plainview Healthcare Partners · Lynbrook, NY
2015 – 2017
Education
Toolkit