Accidental Landlord

Pipeline Steps

Phase 1: Market Setup

1. Use search_locations to get the parcl_id for the target MSA.
2. Confirm the market has both sales and rental data.

Phase 2: Data Pulls

Pull two datasets from property_events:

For-sale listing events (lookback window):

• event_names: ["LISTED_SALE", "LISTING_PRICE_CHANGE", "RELISTED"]
• start_date: 4-5 months before today
• end_date: ~45 days before today
• property_types: ["SINGLE_FAMILY"]
• include_property_details: true
• limit: 20000

Rental listing events (lookahead window):

• event_names: ["LISTED_RENT", "RENTAL_PRICE_CHANGE"]
• start_date: ~45 days before today
• end_date: today
• property_types: ["SINGLE_FAMILY"]
• include_property_details: true
• limit: 20000

Enrichment pulls:

• motivated_seller_properties: top 500 by motivated_seller_index_value. Join to AL leads on parcl_property_id.
• motivated_renter_properties: top 500 by motivated_renter_index_value. Join to AL leads on parcl_property_id.

Phase 3: 110-Day Rolling Window Methodology

For each analysis date (sample every 15 days across the available range):

1. Define windows: Lookback = analysis_date minus 50 days → analysis_date. Lookahead = analysis_date → analysis_date plus 60 days.
2. Find transitions: Properties with for-sale events in lookback AND rental events in lookahead.
3. Same-owner filter: Compare event_entity_owner_name across sale and rental events. Keep only matches (including both-null as a match).
4. Deduplicate: Keep the earliest transition per parcl_property_id.
5. Outlier filtering:
   • Remove rents < $500/mo or > $10,000/mo
   • Remove sale prices < $50,000
   • Compute gross yield = (rental_price x 12) / sale_price x 100
   • Remove yields < 2% or > 18%

Phase 4: Enrichment Joins

Left-join AL leads to motivated seller and motivated renter data on parcl_property_id:

• Adds: motivated_seller_index_value, label, days_on_market, num_price_cuts, agent contact info
• Adds: motivated_renter_index_value, label, rental DOM, num_rent_reductions, property manager info

Phase 5: Analysis

Compute and surface:

• KPI metrics: total leads, median list price, median rent, median/mean gross yield
• Geographic distribution: top ZIPs, top cities (bar charts)
• Owner-occupied flag: strongest PM pitch signal
• Enrichment overlaps: leads appearing in both motivated seller and renter data (double-signal)
• Yield distribution: sort leads by gross yield for actionable table ordering

Phase 6: Dashboard Output

Build a single-file HTML dashboard. No external dependencies.

Structure:

• Header: Market name, date, total leads, "Parcl Labs MCP" branding
• KPI strip: Total leads, median list price, median rent, median gross yield, mean gross yield
• Pitch angle insight box: Gradient background card explaining why accidental landlords are ideal PM prospects
• Distribution charts: Top ZIPs and top cities as horizontal bar charts
• Lead table: Searchable, sortable table with columns: Address, Bed/Bath, Sq Ft, Year, Sale Price, Rent, Gross Yield, Owner-Occupied flag, Signals (motivated seller/renter badges). Sort by gross yield descending.
• Methodology section: Explain the 110-day window, same-owner filter, and outlier exclusions.

Design:

• Color palette: Blue (#2871CC), Orange (#E8841A), Dark Green (#1A2E1F)
• Clean, modern, minimal: white cards, subtle borders, sticky header
• Table max-height with scroll, sticky thead
• Search input for client-side row filtering

Reference Output

See examples/ for the Houston MSA reference implementation:

• houston-dashboard.html: complete dashboard (1,464 leads)
• houston-timeseries.csv: monthly AL rate data
• houston-by-zip.csv: ZIP-level granular analysis
• houston-timeseries-chart.png / houston-by-zip-chart.png: visualizations

See reference/houston-methodology.md for the exact reproduction methodology with validation against published benchmarks.

Display Names

Never display raw parcl_id values in user-facing output. Always resolve to human-readable names:

• ZIP parcl_ids to actual ZIP codes and town names (e.g., 5452730 to "East Hampton (11937)")
• MSA parcl_ids to metro area names (e.g., 2900417 to "Tampa-St. Petersburg-Clearwater, FL")
• Investor IDs to standardized names from search_investors

Note: some MCP endpoints return IDs as floats with .0 suffix (e.g., 5452730.0). Strip the .0 before any lookup or display.

Quality Checks

Before delivering:

• All KPI values match the underlying data
• No outlier records in the table (check for implausible rents, yields)
• Search/filter works on the table
• Methodology section accurately describes what was done
• Gross yield calculation is correct: (rent x 12) / sale_price x 100
• Owner-occupied flag correctly sourced from property metadata
• Motivated seller/renter badges only appear when enrichment data exists for that property

Adapting to Other Markets

This pipeline works for any MSA. High-value target markets (from published research): Houston, Dallas, Phoenix, Tampa, Atlanta, Charlotte. These six metros hold 36.8% of all large institutional SFR holdings and are where accidental landlord formation is accelerating fastest.