NYC Property Data — 5-borough PLUTO pull

What was asked

Give a real-estate operator a concrete, current picture of the highest-value tax lots in New York City, broken out by borough. No enrichment vendors, no scraping tricks — only legal, no-auth public data that anyone can verify.

What the system did

The pipeline hit the NYC Open Data PLUTO dataset (data.cityofnewyork.us Socrata API — Primary Land Use Tax Lot Output, ~860,000 NYC tax lots). It pulled 1,300 records across five boroughs:

• Manhattan: top 500 by total assessed value
• Brooklyn / Queens / Bronx / Staten Island: top 200 each

For every lot it normalized 20 PLUTO columns — BBL (Borough-Block-Lot), address, zipcode, building class, land use, owner, lot and building area, year built, floor count, residential + total units, assessed land, assessed total, primary zoning district, lot dimensions, and latitude/longitude.

All 1,300 rows were then ingested into Postgres as scraped_products rows under marketplace='nyc_pluto' — the raw 20-column JSON rides in the raw_json JSONB column, so any PLUTO field is queryable:

SELECT raw_json->>'address', (raw_json->>'assesstot')::bigint AS usd
FROM scraped_products
WHERE marketplace = 'nyc_pluto'
  AND category = 'Manhattan'
  AND raw_json->>'bldgclass' LIKE 'O%'
ORDER BY usd DESC
LIMIT 20;

Headline numbers

• 1,300 tax lots across 5 boroughs
• $142.8B total assessed value
• 214,757 residential units represented
• Manhattan outweighs every other borough combined — $92.5B across the top 500 MN lots vs $50.3B across the other four boroughs’ top 200 each
• Queens has the highest average top-lot assessment at $277.7M — skewed by a single $8.5B outlier (likely LaGuardia / JFK airport land)

What a pilot engagement gets on top of this

This sample stops at the raw data + a written summary. A real retainer adds:

1. Target-list filtering — e.g. “ground-floor commercial lots in Downtown Brooklyn built before 1950, owned by LLCs, zoned C6-1, with at least 8 floors” — all expressible against raw_json + the typed columns.
2. Owner contact enrichment — resolving LLC owner names to mailing addresses + principal contacts via separate public-record sources.
3. Monthly refresh — scheduled pulls against PLUTO + diff reports when lots change hands or assessments shift.
4. Slide/PDF export — the REPORT.md can render to PDF or Google Slides for client-facing decks.

Caveats

• PLUTO is updated on a ~quarterly cadence by NYC DCP. The “generated on” timestamp reflects the pull date, not the assessment effective date.
• assesstot is the NYC Department of Finance’s assessed value for tax purposes — it is not market value. On commercial lots it tends to sit 30–60% below market.
• Rows with a $0 assesstot (tax-exempt lots, some government property) are excluded from this pull — the $order=assesstot DESC sort pushes them to the end.