Research Documents

These research documents were produced during project planning. They contain detailed analysis of data sources, automation strategies, modeling approaches, and metric definitions. They serve as primary source material for the project.

Research Documents

Automatable Public Data Sources for U.S. Residential Homebuilding Reference

Comprehensive catalog of every federal and government-adjacent data source relevant to a national housing dashboard. Covers the full federal backbone (BPS, NRC/NRS, HMDA, FHFA HPI, FRED, BLS, HUD-USPS) with endpoint URLs, file formats, cadences, geography granularity, and automation difficulty scores (1-5). Establishes the two-tier data architecture concept and proposes a normalization schema.

Key topics covered:

  • Federal backbone source catalog with endpoint patterns
  • Two-tier architecture: federal backbone + local systems
  • CBSA geography dimension (vintage-aware delineations)
  • Metric-to-source mapping with replicability assessment
  • Warehouse schema recommendation
  • CBSA "data availability card" template

Read full document →

Nationalizing the Houston New-Home Dashboard Reference

Detailed connector playbook for building automated data ingestion from federal and local government sources. Includes per-connector ingestion steps (BPS, SOC/NRC/NRS, HMDA, ArcGIS, Socrata, Accela), SQL DDL for all tables, Mermaid ETL diagrams, and a prioritized national rollout plan across top-25 CBSAs.

Key topics covered:

  • Houston presentation metric inventory (60+ metrics cataloged)
  • Per-connector automation playbooks (BPS, HMDA, ArcGIS, Socrata, Accela, Tyler, OpenGov)
  • Permit-to-start and finished-vacant modeling methods
  • SQL DDL for permits, lifecycle, and closings tables
  • Sample SQL queries for Q4 starts, completions, UC stock, finished vacant
  • National rollout prioritization heuristic
  • CBSA data availability card template (prefilled for Houston)
  • Release timeline Gantt (BPS 17th workday, NRC/NRS 12th workday, etc.)

Read full document →

Reproducing the Houston CBAS Metrics Nationally Reference

Detailed metric-by-metric reproduction plan starting from the Houston CBAS presentation. Inventories every metric from the deck (70+ items), maps each to the best public data sources, assesses national replicability, and provides per-metric reproduction recipes with SQL query examples.

Key topics covered:

  • Complete Houston deck metric inventory (demographics, employment, resale, new home, lots, pricing, focus areas)
  • Ambiguity analysis: what constitutes a "start", "closing", "finished vacant", etc.
  • Per-metric reproduction recipes (permits, starts, UC, completions, closings, finished vacant)
  • Three estimation methods: federal-first, lifecycle-fusion, USPS vacancy proxy
  • Gap analysis: what's replicable vs proprietary
  • Houston-specific source inference (what CBAS likely used)
  • SQL DDL for canonical tables (dim_jurisdiction, fact_permit, fact_inspection_event, fact_unit_lifecycle, etc.)
  • Sample SQL for Q4 metrics, UC stock, finished vacant inventory, months' supply

Read full document →

Design Documents

Houston MVP Design Implemented

Full design document for the Houston MVP. Covers architecture decisions (Supabase, Python ingest, Next.js dashboard, monorepo), complete database schema DDL, dashboard page specifications (6 pages with KPI definitions), the permit-lag pipeline model, and 3-phase MVP rollout plan.

Read design doc →

Phase 1 Implementation Plan Implemented

Detailed implementation plan for Phase 1 (federal backbone + modeled pipeline). 11 tasks covering project scaffolding, Supabase migrations, Python connectors (BPS, FHFA HPI), pipeline model, Click CLI, and Next.js dashboard scaffold. All tasks completed and merged.

Read plan →

Phase 2 Design Implemented

Design document for Phase 2 (employment, population, HMDA, FRED). Covers new connectors, database migrations, CLI commands, dashboard pages, and the HMDA closings integration into the pipeline model.

Read design doc →

Phase 2 Implementation Plan Implemented

Detailed implementation plan for Phase 2. Covers BLS employment connector, Census PEP population connector, FRED API connector, HMDA mortgage origination connector, four new SQL migrations, and three new dashboard pages (employment, demographics, prices).

Read plan →