← All work · Self-initiated demonstration on public data
Case study · Market analysis
Florida Market-Analysis Pipeline
Pulls official Census figures for every Florida county, checks them, and turns them into trustworthy market numbers, with the fine print kept intact.
The problem
Market analysis for real estate and insurance leans on county population and income figures. Raw Census data is easy to get wrong: blank-value codes counted as real numbers, error margins dropped, and county totals that don't add up to the state.
What it does
- Downloads American Community Survey data for all 67 Florida counties and checks it before saving anything.
- Can be re-run safely at any time without creating duplicates.
- Keeps the margin of error with every figure and flags estimates that are too uncertain to rely on.
- Automatically checks that the 67 counties add up to the official state total.
- Defines each market metric once, with its caveats, so everyone uses the same numbers.
What you get
- Reliable county-level figures ready for market maps and reports.
- Clear flags on numbers that shouldn't be trusted.
- Easy to extend to other states by changing one setting.
By the numbers
67Florida counties plus the state total, with margins of errorSource: census-pipeline README
Σ 67 = statecounties checked against the official state totalSource: census-pipeline README
CV > 0.30cut-off that flags unreliable income estimatesSource: census-pipeline README
Built with: Python, Census API, DuckDB, dbt
Contact
Tell me about your data.
Send a short note about your area, your data and what you need. I reply by email, usually within one business day, or you can book a call directly.