How Stratara reads U.S. markets

The principle

The product is the methodology. CoStar publishes data behind a $30,000/yr paywall. Green Street publishes opinions in PDFs. Stratara publishes both — transparently — at mid-market prices. See how we stack up on the Compare page.

Real estate is downstream of demand drivers. Stratara is built on the premise that by the time a market reprices, the jobs, people, capital, and infrastructure that drove the move have already been visible in public datasets for months. This platform aggregates those signals — not just price history — so you see the upstream picture before it shows up in the comp sheet.

Data sources

All sources are public. Most are primary government surveys or regulatory filings. We re-ingest on each source’s native cadence and cache aggressively to keep page loads fast.

Demographics & Population

Economy & Labor

Housing & Real Estate

Health & Social

Environment & Risk

Education

Infrastructure & Connectivity

Crime & Policy

Composite indices & spoke scores

Every metro and county gets a single top-line composite score (0–100). It is the equal-weighted average of up to six substantiated composite indices — Investment, Affordability, Market Heat, Growth Momentum, Quality of Life, and Rental Strength. There are no fixed per-index weights: each available index counts equally toward the mean.

The average is null-disciplined. We average only the indices we can actually substantiate for a geography — a missing index is dropped from the mean, never defaulted to a neutral “50.” If none of the six are available, the composite is left blank rather than fabricated. A market substantiated on fewer indices will reflect that directly as a narrower composite — missing indices are dropped from the mean and never defaulted to a neutral score, so a two-index market’s composite is the mean of those two indices only.

The 8-spoke diagnostic radar

Beneath the top-line score, each market is profiled on an eight-spoke radar — the deeper diagnostic view shown on every metro and county detail page. The eight spokes are Investment, Market Heat, Growth, People, Housing, Quality of Life, Safety, and Climate & Risk. Each spoke is a blend of three sub-composites, and each sub-composite is built from five underlying metrics. A spoke renders as null — never defaulted to 50 — when its inputs are missing for a geography; sparse spokes are flagged visually rather than filled in.

Advanced Composites

Alongside the top-line composite and the eight-spoke radar, Stratara computes a set of standalone composite indices that power the Rankings leaderboard and Investment Analysis tab. These do not feed the top-line composite score but are available for screening and sorting across all 200+ metros and 3,000+ counties. Each is a weighted blend with proportional redistribution: a missing input is dropped and the remaining weights are renormalized to carry the full signal, but a composite returns null whenever every input is missing or less than half of the intended signal weight is available — never a synthetic fill.

Fair-value cap rate model

The fair-value cap rate is Stratara’s read of where a CRE market should price, given its fundamentals. Compared against the observed cap rate, it produces a verdict — CHEAP, FAIR, or RICH — plus a basis-point spread you can quote in a memo.

• Base yield — current 10-year Treasury pulled live from FRED, recomputed daily. Never hardcoded. Institutional convention: all CRE cap rate spreads are quoted against the 10-year.
• Illiquidity premium — 100bps above the risk-free rate, representing the normal-cycle cost of real estate’s lack of daily liquidity. Calibrated against CBRE Investment Management’s 342bps long-run spread (1991–2019) minus observed historical growth credits and risk premia.
• Regional risk premium — derived from Stratara’s economic, risk, and legal spoke composites. Ranges 0.50–2.50%: institutional gateway markets earn compression; tertiary and high-vacancy markets pay full premium. Reflects employment diversification, vacancy, and regulatory environment per Plazzi/Torous/Valkanov 2010.
• Growth credit — composite of population growth (Census/IRS), job growth (BLS QCEW), and MF rent growth (Census ACS median gross rent + HUD Fair Market Rents). Capped at 150bps to prevent ZIRP-era overshoot. Supply-pipeline haircut reduces credit by 25–100bps when forward 24-month deliveries exceed 3–7% of existing stock.
• Climate load — piecewise-linear load from FEMA NRI expected annual loss (EAL) percentile. Ranges 0–200bps: low-risk markets (NRI <40) pay 0–25bps; high-risk markets (NRI >80, e.g. Tampa, Miami, Houston) pay 100–150bps. Calibrated from CBRE’s finding that rising insurance costs drove a 3.6% national MF value decline since Q4 2019, with South-Central at −7.8% and Florida at −6.8%. Academic basis: Bernstein/Gustafson/Lewis 2019 (JFE), Goldsmith-Pinkham et al. 2023 (RFS), Federal Reserve FEDS Note Sept 2025.
• Observed-side reference — the verdict compares Stratara’s per-market fair cap rate against a national MF reference cap rate (10-yr Treasury plus a national MF spread constant). This means the panel reads each market’s fundamentals against national pricing, not that market’s own transaction history. Per-market observed cap rate data will replace the national reference as coverage expands.

The verdict is the spread between fair and observed. Fundamentals implying a cap rate ≥50bps wide of the national reference screen CHEAP; ≤−50bps screens RICH. FAIR is everything in between. Climate-exposed markets will frequently show RICH — this is intentional: the model prices the forward risk the market has not yet fully capitalized.

What we don’t show

Equally important is what’s not on the page. Stratara’s policy is empty-state over fabrication:

• If we don’t have annual history for a geography, the historical chart shows an empty card — not a synthesized trend.
• If a composite input is missing, the spoke may be null — not defaulted to 50. Missing data degrades the data confidence score visibly.
• If FRED is unreachable, yield-dependent fields show “—” — not the last-known stale value.
• If a source hasn’t been ingested yet or the latest vintage isn’t available, the field is blank.
• If a metric is suppressed at the county level due to small sample size (common in ACS and CDC datasets), we display the suppression indicator rather than the censored value.

Update cadence

• Daily: 10-yr Treasury & Fed Funds (FRED) — all yield-dependent fields recomputed on each pull
• Monthly: BLS LAUS unemployment, Zillow ZHVI/ZORI, FHFA HPI, NOAA storm events
• Quarterly: BLS QCEW employment by sector, Census building permits, FCC broadband availability, EIA electricity prices
• Annual: Census ACS demographics & housing, IRS migration flows, BEA GDP & personal income, FEMA NRI, FBI crime, CDC PLACES, CDC SVI, CDC overdose mortality, EPA AQS air quality, EPA EJScreen, EPA walkability, NOAA climate normals, HUD FMR & CHAS, NCES IPEDS & CCD, FDIC bank deposits, CMS Medicare, DOE energy, BLS OEWS, DOT BTS airports, USDA rural, Eviction Lab, MIT elections, ACS housing & commute extras
• On change / periodic: USAspending.gov contract awards, Inside Airbnb STR listings

Each metric card in the app displays the data vintage year so you can see exactly which ACS release or which FEMA NRI version is behind a given number.

Limitations & attributions

• Commercial CRE sub-market data (multifamily occupancy, office vacancy, retail/industrial absorption, CRE transaction cap rates) depends on commercial data licensing from CoStar, MSCI/RCA, or similar. These fields are not currently in Stratara.
• Small-area suppression: Many federal datasets suppress county-level figures for small populations (typically n<20). Stratara displays the suppression indicator rather than fabricating a value.
• Geographic unit mismatch: Some sources publish at ZIP code, census tract, or metropolitan division — not county. We aggregate upward or map to the nearest containing geography using Census TIGER/Line crosswalks.
• Data is for informational purposes only. Stratara is not investment advice.
• We update this page when the methodology changes materially.

Third-Party Attributions

• Eviction Lab data is provided under a Creative Commons Attribution 4.0 International License (CC BY 4.0). Princeton University Eviction Lab, evictionlab.org.
• All other sources are U.S. federal government publications in the public domain, or publicly licensed under open data terms.