CBAS Parity Analysis

License a data feed from the local REALTOR association (HAR in Houston) via the RESO Web API standard. Alternatively, use a national aggregator like CoreLogic or ATTOM.

Unlocks:

• Resale snapshot — annual/monthly sales, prices, days on market, active listings, sale-to-list ratio
• Resale closings vs active inventory time series
• Median new-home price + price/SF — via new-construction filter on MLS listings

Residual gap: MLS new-construction tagging varies by market. List price ≠ base price (CBAS tracks base prices before incentives).

Obtain bulk permit + inspection event data from local jurisdictions via TPIA/FOIA requests or negotiate API access to platforms like Accela, Tyler, or OpenGov. For many metros, open-data portals (ArcGIS, Socrata) provide this data for free.

Unlocks:

• Quarterly/annual starts — foundation inspection date = start per Census SOC definition
• Under construction stock — started but not yet completed
• Completions — CO / final inspection = completion
• Finished vacant inventory — completed minus sold (combined with recorder/HMDA data)
• Finished vacant months’ supply (derived)
• Starts/closings by geography — once geocoded, can aggregate to any polygon
• Builder attribution — permit applicant/contractor field as proxy (~60–70% accurate; needs normalization)

Residual gap: Inspection event definitions vary by jurisdiction (Houston uses “Foundation”; another metro may use “Footing” or “Slab”). Normalization required per metro.

Commercial property data providers like ATTOM Data and CoreLogic maintain national deed/transaction databases that consolidate county recorder data. Alternatively, file TPIA requests directly with county clerk offices for bulk deed exports.

Unlocks:

• Total new-home closings (financed + cash) — HMDA covers ~70% (mortgaged); recorder adds the remaining cash transactions
• Quarterly closing granularity — recorder has daily recording dates vs HMDA’s annual bulk release
• Sale date linkage for finished-vacant computation — “completed but not yet sold” requires knowing when each unit sold

Residual gap: Texas does not mandate sale price disclosure on deeds, so Houston recorder data often has blank price fields. Must cross-reference with HMDA or MLS for price. Some cash transfers lack clear “new construction” flags.

Most county assessor offices publish parcel GIS data with lot boundaries, frontage/area dimensions, land use codes, and ownership. HCAD (Harris County) offers quarterly GIS downloads. This is nearly always free or very cheap.

Unlocks:

• Lot-size segmentation — parcel frontage/area fields enable the 40’/50’/60’ lot-width bands CBAS uses
• Starts/closings by lot size — spatial join permits to parcels, then bin by lot width
• Finished vacant by lot size — same join applied to inventory computation
• Parcel-level joins for all metrics — the backbone for connecting permits, deeds, and inspections to physical lots

Residual gap: ~5% of parcels have missing/incomplete frontage fields in some counties. Quality varies.

Monthly satellite imagery (3m resolution from Planet Labs, or higher from Maxar/Nearmap) enables detection of ground disturbance, site clearing, and construction activity through temporal change analysis. This is the closest substitute for CBAS’s field verification of lot pipeline status.

Unlocks:

• VDL detection — identify cleared/graded lots that haven’t started vertical construction
• Future lots with active site work — detect clearing, grading, utility trenching, paving via month-over-month image comparison
• Construction progress — supplement permit data with visual confirmation of starts and completions

Residual gap: Satellite can detect visible surface changes but misses underground utilities, ownership intent, liens/encumbrances, and lot “readiness to build” status. False positives from non-residential earthwork. Monthly cadence means 2–4 week lag.

Maintain a daily/weekly scrape of builder websites for floorplan catalogs (base prices, specs, community assignments). Supplement with direct data feeds negotiated with major public builders (Lennar, Pulte, Toll Brothers, etc.). The top 10–20 builders typically cover ~60% of metro production volume.

Unlocks:

• Floorplan QoQ price direction — detect base price changes across the catalog
• Cheapest/most expensive floorplan profiles — rank active plans by current base price
• Starts/closings by base price band — link lifecycle events to floorplan base prices
• Finished vacant by price band

Residual gap: Base price ≠ list price. CBAS distinguishes genuine price cuts from incentive changes — scraped data conflates them (~20–30% of observed “price changes” are actually incentive shifts). Smaller builders often don’t list base prices publicly. Plan ID normalization when builders rename/retire plans is manual.

Define communities from plat polygons + permit text field parsing + builder roster curation. Define submarkets/corridors as ZIP code clusters, county subareas, or planning-department market areas. Validate against 2+ years of permit/sales data. This is labor, not a data license — someone has to curate and maintain the reference tables.

Unlocks:

• Top communities by starts/closings — aggregate permits to community polygons, rank
• Top builders by market share — permit applicant + manual builder roster
• All corridor/submarket breakdowns — once polygons defined, any geocoded metric can be sliced by submarket
• Submarket maps with community locations

Residual gap: CBAS submarket/corridor boundaries are proprietary and informed by market knowledge (traffic patterns, school districts, builder focus areas). Our ZIP-based or plat-based approximations will differ. Community names are marketing constructs that don’t always match legal subdivision boundaries. Builder attribution from permit applicant fields is ~60–70% accurate (GC ≠ builder).

Contract 2–4 field specialists per metro to sample 20–30% of VDL inventory quarterly and validate lot status, site-work conditions, and community boundaries. This is what CBAS does — “visually verified” is their core differentiator.

Unlocks:

• VDL field verification — confirm that “appears buildable” lots are genuinely ready to build (utilities connected, no liens, owner willing)
• Site-work condition taxonomy — standardize lot status beyond what satellite/permits show (raw, cleared, utilities, paving)
• Community/builder ground truth — validate which subdivisions are active, which builders are present, model home counts

What remains: Even with field sampling, ~10–15% of lot inventory status changes between sampling visits. CBAS likely has higher cadence and deeper builder relationships built over years.