first-market-county-data

Research and locate first-to-market distress data sources for real estate wholesaling. Use when the user needs to find WHERE to pull county-level distress lists (probates, foreclosures, tax sales, lie

SKILL.md

GitHub

First-to-Market County Data Research

This skill helps real estate investors find the exact offices, portals, and processes to pull first-to-market distress data for any U.S. county.

Data Priority Pyramid Context

This skill focuses on Tier 1 - First-to-Market Data within the Data Priority Pyramid:

Pyramid Level Data Source Examples
Tier 1 First-to-market county data (THIS SKILL) Probate, foreclosure, tax sale pulled directly from county
Tier 2 Nationwide aggregated data DataSift, Propsumer, Batch
Tier 3 AI-enhanced data DataSift AI products

Always prioritize Tier 1 data sources before moving to Tier 2 or Tier 3.

What is First-to-Market Data?

First-to-market data means pulling distress records (probates, foreclosures, tax sales, etc.) the day or week they become available—before competitors. This data often requires:

  • Physical county visits or special access
  • Signed affidavits or license verification
  • FOIA/public records requests
  • Understanding of county-specific regulations

When to Use This Skill

Use when the user asks to:

  • Find where to pull distress data for a county
  • Research county recorder/clerk offices
  • Locate probate, foreclosure, tax sale, or lien sources
  • Run the "deep research prompt" for county data
  • Set up first-to-market data pulling processes

Core Workflow

  1. Collect target location: Get county name and state from user
  2. Select data types: Determine which distress lists are needed — default to full Priority A + B + C coverage unless the user asks for a subset
  3. Sweep the cross-cutting sources listed in the "Always-Check Cross-Cutting Sources" section below before falling back to data-type-specific prompts. Many counties have statewide aggregators, court journals, and alternate-auction pipelines that get missed if you only ask "where does the county file X?"
  4. Run research prompts: Execute prompts from references/research-prompts.md
  5. Populate multi-URL portal fields: For every data row, attempt to list the primary portal + at least one cross-validation / backup URL in the Portal URL field (newline-separated). A single URL is only acceptable when you've confirmed no alternate exists. See "Multi-URL Portal Convention" below.
  6. Cross-check against any prior XLSX the user has for nearby counties in the same state. State-level infrastructure (judicial vs. non-judicial, tax lien vs. tax deed, statewide aggregator URL, court-journal pattern) repeats across counties — the previous report is a checklist. Missing entries that appear on a sibling-county report are a red flag.
  7. Compile results: Build an input JSON following the schema in scripts/build_spreadsheet.py
  8. Deliver XLSX via the template: Run python scripts/build_spreadsheet.py input.json output.xlsx — this is the only sanctioned output format. Do not invent alternate layouts. See references/spreadsheet-template.md for the full spec.

The XLSX file is the deliverable. Always write it to the project's output/ftm_county_research/ folder (create the folder if it doesn't exist; never put it at the top level of output/). Link it prominently at the top of your chat response, and include a short summary + recommended action order underneath. Do not stop at a markdown-only answer or a CSV.

Output file naming: {County}_County_First_Market_Data_Sources.xlsx — overwrite prior versions rather than date-stamping, unless the user explicitly requests versioned snapshots.

Always-Check Cross-Cutting Sources

Most counties have distress-data sources that don't map cleanly to a single data type. These get missed when you only run the per-data-type prompts. Sweep for all of these before finalizing the report. If a category doesn't apply to this state/county, note that explicitly in the Notes field of a related row rather than silently omitting it.

Category What to look for Why it matters
Statewide public-notice aggregator Search {state} public notices — most states have one (OH: publicnoticesohio.com, TN: tnpublicnotice.com, FL: floridapublicnotices.com, TX: texaspressassociation.com/publicnotice). Same ASP.NET platform family across states → one scraper covers many counties. Often the cleanest legal source for foreclosure + probate notices. List as its own Priority A row.
State-government notice site Search {state}.gov public notice (OH: publicnotice.ohio.gov). Separate from the newspaper aggregator. State agency notices, hearings, forfeitures. Second URL on the statewide aggregator row.
County court journal / legal newspaper Search {county} legal news or {county} daily record. Designated by statute as court journal of record (e.g., Akron Legal News, Daily Legal News Cleveland, Daily Report Atlanta, New York Law Journal). Publishes foreclosures, probate, tax delinquent lists, sheriff sale notices. Often the free source for the annual certified delinquent list. List as its own Priority A row.
Forfeited / excess-land sale Properties that fail multiple tax-sale rounds are forfeited to the state and auctioned by the Auditor/Commissioner. Secondary deep-distress inventory, lower competition. Separate Priority A row — do not merge with tax sale.
Land bank demolition pipeline Search {county} land bank — most rust-belt / post-industrial counties have one. Publishes demolition target addresses before teardown — first-to-market signal on structurally-distressed vacants. Add as a URL on the Condemned row (and in Notes).
Open Data portal (Socrata / ArcGIS REST) Search data.{city}.gov or {city} open data. Primary cities in data-friendly states (OH Cleveland, IL Chicago, NY NYC, CA SF) publish code violations / permits / lead certs via REST API. Easiest possible scrape target — replaces FOIA for whichever data types the city publishes. List the specific dataset URL, not just the portal root.
Recorder document-type searches Multiple lien types (mechanic's, federal tax, state tax, Medicaid, HOA, child support, judgment) are typically all filed at the same county recorder portal, filtered by document type. The portal URL repeats across 5-8 rows — that's expected, and correct. Difference is the document-type filter in the Notes field.
Fraud-alert / property-alert subscription Many recorders offer free property-fraud alerts by email when a deed is filed against a parcel. Passive watchlist for target-parcel distress signals. Note as a capability in the recorder row.
State attorney-general registries State AGs often hold secondary indexes (state tax liens, Medicaid recovery, charity registrations). Use as a cross-validation URL on the relevant rows, not primary.
RealAuction vs. county-native auction sites Ohio sheriff sales migrated to {county}.sheriffsaleauction.ohio.gov; some counties keep a sheriff.{county}.gov/sheriff-sales native page with additional info (bidding rules, results archive). Both URLs belong on the sheriff-sale row. RealAuction sometimes 403's bot UAs — county-native page is the fallback.

State-specific templates

Before researching a county, search for a prior county report from the same state. If one exists, use its non-county-specific fields (statewide aggregator, state tax-lien process, forfeited-land statute, state sheriff-sale vendor, ODNR/EPA-equivalent) as the starting point. You're researching the county-specific fields, not re-deriving the state infrastructure from scratch.

Multi-URL Portal Convention

The Portal URL field (column F) accepts newline-separated URLs. Default to listing 2-3 URLs per row:

  1. Primary portal — the one you'd scrape first
  2. Cross-validation source — a different system that carries the same data (e.g., court journal publishing foreclosure notices, state aggregator filtering to the county)
  3. Backup / alternate — fallback when primary is rate-limited, blocked, or paywalled

Only list a single URL when you have explicitly verified (WebFetch / WebSearch) that no alternate exists. If the only second URL is a thin/marketing page, leave it out rather than padding.

Example (good):

https://summit.sheriffsaleauction.ohio.gov
https://sheriff.summitoh.net/pages/Sheriff-Sales.html
https://www.akronlegalnews.com/notices/sheriff_sales

Example (bad — single URL with no cross-validation):

https://summit.sheriffsaleauction.ohio.gov

(Unacceptable unless you've confirmed no county-native page and no court-journal republishing exists.)

Data Types Available

Priority A - Core Lists (Big Vexation)

Data Type Description
Probate/Heirship Estates, unknown heirs, heirship determinations
Foreclosure/Auction Lis pendens, in-rem foreclosures, auction lists
Tax Sale Properties going to tax auction
Tax Delinquency Properties with unpaid property taxes

Priority B - Standard Lists

Data Type Description
Code Violations Municipal code enforcement fines
Condemned Structures Unsafe/condemned building registers
Mechanic's Liens Contractor liens against property
IRS/State Tax Liens Federal and state tax liens on real estate

Priority C - Extended Lists

Data Type Description
HOA/Condo Liens Assessment liens from associations
Utility Shut-offs Water/electric disconnection lists
Building Permits Open/expired residential permits
Environmental Citations Mold, asbestos, lead hazard citations
Storm/Fire Damage Incident reports from Fire Marshal
Medicaid Recovery Liens Estate recovery liens
Multiple-Eviction Landlords Properties with 2+ eviction filings
Child Support Liens Judgment liens on real property
Quiet-Title Suits Post-tax-deed quiet title lawsuits
Sinkhole/Subsidence Geological survey claims data

Research Prompt Format

For each data type, use this prompt structure:

Act as a public-records researcher. For {{COUNTY}} County, {{STATE}}, 
identify the office that [records/maintains/handles] [DATA_TYPE].

Return in one markdown table row:
• Office/division name
• Street address & phone #
• Online index/portal link (if any)
• Low / Medium / High estimate of difficulty to bulk-download the data

See references/research-prompts.md for complete copy-paste prompts for all 16+ data types.

Output Format — Canonical Template

Required: 10-column XLSX built via scripts/build_spreadsheet.py. Full spec in references/spreadsheet-template.md.

Columns (A–J, fixed order):

# Column Contents
A Priority A / B / C
B Data Type e.g., "Probate / Heirship", "Lis Pendens / Foreclosure"
C Office / Division Full office name
D Address Street + city/state/zip (multi-line OK)
E Phone Formatted with area code
F Portal URL Primary online search URL
G FOIA / Contact Email FOIA form URL or contact email
H Difficulty Low / Low-Medium / Medium / High
I Notes Access quirks, legal caveats, scraping hot-buttons
J Data Freshness e.g., "Daily filings", "Weekly auction schedule", "Annual PDF"

Includes a second Legend & Notes sheet with priority/difficulty legends and county-specific context.

Difficulty Rating Guide

Rating Meaning
Low Online portal with bulk export; minimal barriers
Low-Medium Online searchable but bulk export limited; may need account or FOIA
Medium Account registration, limited export, or in-person pickup required
High FOIA request, physical visit, affidavit, or paywall required

Example Usage

User request: "Find where to pull first-to-market data for Knox County, Tennessee"

Response workflow:

  1. Run research prompts for Priority A data types (probate, foreclosure, tax sale, tax delinquency)
  2. Compile findings into spreadsheet
  3. Note any special requirements (e.g., Knox County requires physical visit + affidavit for tax sale data)
  4. Deliver CSV/XLSX with all source details

Important Notes

  • Some counties have strict anti-predatory regulations requiring in-person visits
  • Data freshness varies: probates may be weekly, auctions may be monthly
  • Always verify current requirements—county processes change frequently
  • Water shut-off lists are highly valuable but often difficult to obtain
  • First-to-market advantage diminishes if data is also available through PropStream/BatchLeads

Reference Files

  • references/research-prompts.md — Complete copy-paste prompts for all data types
  • references/common-offices.md — Typical office names by data type
  • references/spreadsheet-template.md — Canonical XLSX output spec (columns, colors, sheet structure)
  • scripts/build_spreadsheet.py — Template builder; takes input JSON, produces the XLSX