Ohio sheriff sale dashboard hero

Real Estate Investor Dashboard

Playwright scraper, Supabase, and a Next.js dashboard that replaced 88 tabs and a spreadsheet with one filterable view.

Client Overview

Real estate investor specializing in foreclosures

Lean acquisition operation targeting sheriff sale properties before they hit the open market. A VA was manually checking 88 county sites and logging results in a spreadsheet. The process was inconsistent, error-prone, and hard to scale.

Problem

88 individual county sites

  • Each county has its own site with a different layout

  • Three distinct patterns: realforeclose.com subdomains, county portals, and JavaScript-rendered sites

  • VA spent 3 to 4 hours per day clicking through county sites, still missed listings

Solution

Playwright scraper + Supabase database + Next.js dashboard

  • 88-county config file drives all scrape targets, one row per county, no code changes to add counties

  • Three scraper strategies unified behind a single orchestration loop: httpx + BeautifulSoup for realforeclose.com, custom selectors for county portals, Playwright for JS-gated sites

  • Listings upsert on (county_slug, case_number), with scrape_runs logged per county for health monitoring

  • Auto-enrichment runs after each scrape: geocoding, Street View, Zillow via Apify, with Firecrawl as URL discovery fallback

  • Next.js dashboard reads from Supabase: listings, county health, saved lists, and batch enrichment retry from the error log

Tech Stack

Modal runs the scraper and enrichment API. Listings land in Supabase. The Next.js dashboard is read-only against the database.

Modal

Scraper runtime and FastAPI enrichment endpoint

Playwright

Headless Chromium for JS-rendered county sites

httpx + BeautifulSoup

Static HTML county sites and realforeclose.com

Supabase

PostgreSQL with RLS, listings, counties, scrape_runs, enrichment_errors

Next.js

App Router, TypeScript, Tailwind, shadcn/ui

Google Maps

Geocoding and Street View

Apify

Zillow property data actor

Firecrawl

Zillow URL discovery fallback

Outcomes

  • 79 of 88 counties in production (89.8%), login-walled counties auto-flagged inactive

  • Daily automated scrape replaces manual VA county checks

  • One dashboard to filter listings by county, sale date, and bid amount instead of jumping between tabs and spreadsheets

  • Property enrichment at scale without per-listing Zillow lookups

Project Walkthrough