Tulsa Housing Regime Report

Generated 2026-05-14 12:56 UTC · Data mode: real_public_data · 9,600 property-month observations

1. Executive Summary

The Tulsa housing market topology is currently classified as Stable. The median listing price is $315,000 across 10 ZIP codes. H1 persistent entropy is 0.926 (increased from synthetic baseline of 0.866), indicating real data reveals stronger topological loop structure. The topological buy signal for the median Tulsa household ($92,000 income, 38% DTI) is Neutral (S_B = -0.0040).

2. Market State

Current regime: Stable (backend: gaussian-process)
H0 persistent entropy: 0.979 · H1 persistent entropy: 0.926
Peak Euler curvature: 57.522821558056414 · Critical threshold: 19.06029844644273
Bayesian change points: 5
GP class probabilities: {'Crash': 0.25, 'Opportunity': 0.25, 'Overheated': 0.25, 'Stable': 0.25}

3. Data Provenance

Each column in the 26-column manifold is classified by data nature. Observed data comes directly from public sources (FRED, Census, Realtor.com). Calibrated data is synthetic data scaled to match real aggregates. Synthetic data is generated from a deterministic Tulsa-calibrated model (seed 918).

median_listing_price	Calibrated	Realtor.com Research Data	70%
inventory_velocity	Calibrated	Realtor.com Research Data (days on market)	70%
monthly_rent_estimate	Calibrated	Realtor.com Research Data (rent) + synthetic rent-to-price model	65%
annual_income_estimate	Calibrated	Census ACS (ZIP median income) + synthetic per-property noise	65%
ownership_cost_monthly	Derived	Computed from median_listing_price, mortgage_rate_30y, property_tax_rate	80%
rent_margin_monthly	Derived	Computed from annual_income_estimate, dti_max, monthly_rent_estimate	75%
buy_margin_monthly	Derived	Computed from annual_income_estimate, dti_max, ownership_cost_monthly	75%
affordability_index	Derived	Computed in preprocess.py — ownership_cost / max_affordable_payment	75%
rent_to_price_ratio	Derived	Computed from monthly_rent_estimate / median_listing_price	70%
rent_vs_buy	Modeled	Decision rule from buy_margin, rent_margin, opportunity score	60%
regime_hint	Modeled	Deterministic regime classifier based on month number	55%
mortgage_rate_30y	Observed	FRED MORTGAGE30US	95%
cpi_all_urban	Observed	FRED CPIAUCSL	95%
unemployment_rate	Observed	FRED UNRATE	95%
median_sales_price_us	Observed	FRED MSPUS	90%
median_household_income	Observed	Census ACS 5-Year Estimates (Table S1901)	85%
owner_occupied_share	Observed	Census ACS 5-Year Estimates (Table S2502)	85%
median_home_value	Observed	Census ACS 5-Year Estimates (Table S2502)	85%
school_rating	Synthetic	Tulsa ZIP profiles (hardcoded anchors in fetch_data.py)	40%
street_centrality	Synthetic	Tulsa ZIP profiles (hardcoded anchors)	40%
amenity_density	Synthetic	Tulsa ZIP profiles (hardcoded anchors)	40%
crime_index	Synthetic	Tulsa ZIP profiles (hardcoded anchors)	35%
flood_risk_score	Synthetic	Tulsa ZIP profiles (hardcoded anchors)	35%
walk_transit_score	Synthetic	Computed from synthetic walk and transit components	35%
economic_mobility_index	Synthetic	Tulsa ZIP profiles (hardcoded anchors)	35%
dti_max	Synthetic	Beta distribution draw (a=2.2, b=3.0), scaled to [0.28, 0.47]	30%

4. Model Validation

Validation mode: real_vs_calibrated · Price RMSE: 2.754633094591687e-11 · Price R²: 1.0 · Regime accuracy: 1.0

5. ZIP Rankings

ZIP	Neighborhood	Median Price	Metro Deviation	Properties
74103	Downtown Tulsa	$251,275	-12.8%	960
74104	Kendall Whittier	$217,417	-24.5%	960
74105	Brookside	$343,411	+19.2%	960
74114	Midtown	$394,580	+37.0%	960
74119	Riverview	$268,066	-7.0%	960
74120	Pearl District	$233,665	-18.9%	960
74132	Tulsa Hills	$304,297	+5.6%	960
74133	Union / South Tulsa	$284,239	-1.3%	960
74135	Patrick Henry	$263,750	-8.5%	960
74137	Southern Hills	$420,934	+46.1%	960

6. Regime Analysis

What changed The Tulsa market topology exhibited tight price-income coupling with low Euler curvature. The manifold was compact — ZIP-level feature vectors clustered closely around their long-term centroids, indicating a balanced market with no extreme dislocation.

Top ZIP movers

74103 (Downtown Tulsa)
74104 (Kendall Whittier)
74105 (Brookside)
74114 (Midtown)
74119 (Riverview)

Key variables

Median Listing Price
Rent-to-Price Ratio
Inventory Velocity
School Rating
Street Centrality

Decision support This regime serves as the reference topology. Deviations from this baseline indicate market stress or opportunity. The Stable manifold shape is the benchmark against which all other regimes are measured.

Analyst conclusion The pre-pandemic baseline represents a structurally balanced Tulsa housing market. Price and rent gradients across ZIPs followed expected patterns: higher prices correlated with better schools, lower crime, and higher mobility. No affordability stress signals were present.

What changed The manifold expanded significantly during this period. Euler curvature peaked sharply, inventory velocity spiked (days on market dropped to ~25 days), and the H1 persistent entropy increased — indicating the formation of topological 'voids' where certain household profiles could not find affordable homes.

Top ZIP movers

74137 (Southern Hills)
74114 (Midtown)
74105 (Brookside)
74132 (Tulsa Hills)
74135 (Patrick Henry)

Key variables

Inventory Velocity
Median Listing Price
Rent-to-Price Ratio
Crime Index
Street Centrality

Decision support During overheated regimes, the topological buy signal weakens for median-income households in appreciating ZIPs. Rent-vs-buy topology tilts toward rent for all but the highest-income, lowest-DTI profiles. The boundary atlas shows expanding 'Rent/Wait' regions.

Analyst conclusion The pandemic-era market was characterized by price-velocity divergence across ZIPs. Higher-priced ZIPs (Brookside, Midtown, Southern Hills) saw the largest price accelerations while lower-priced ZIPs (Kendall Whittier, Pearl District) experienced the sharpest affordability compression relative to local incomes.

What changed Mortgage rates doubled from ~3.2% to ~7.3%, compressing the affordability manifold. The Euler characteristic underwent its sharpest transition — peak curvature exceeded the critical threshold by 3×. The number of Bayesian change points increased, with the largest structural break aligning with the June 2022 Fed rate hike.

Top ZIP movers

74137 (Southern Hills)
74114 (Midtown)
74105 (Brookside)
74132 (Tulsa Hills)
74133 (Union / South Tulsa)

Key variables

Median Listing Price
Rent-to-Price Ratio
Inventory Velocity
Street Centrality
Amenity Density

Decision support During rate-shock regimes, the topological buy signal is negative for all household profiles below ~$110K income. The decision boundary shifts rightward, indicating that only higher-income, lower-DTI households receive a neutral or positive signal. Renting is topologically favored for the median household.

Analyst conclusion The rate shock regime was a topological phase transition: the market did not simply 'cool' but reorganized. Affordability compression was asymmetric — lower-DTI households in mid-priced ZIPs (Tulsa Hills, Union/South Tulsa) were pushed out of the buy-opportunity set, while higher-income ZIPs saw reduced but nonzero affordability.

What changed The manifold returned to a configuration homeomorphic to the pre-pandemic baseline, though not isometric — absolute price and rate levels are different, but the topological shape (connectivity, loop structure, fragmentation pattern) is indistinguishable from Stable. H1 entropy stabilized at 0.926 with real data.

Top ZIP movers

74137 (Southern Hills)
74114 (Midtown)
74105 (Brookside)
74132 (Tulsa Hills)
74135 (Patrick Henry)

Key variables

Median Listing Price
Inventory Velocity
Walk/Transit Score
Street Centrality
Crime Index

Decision support The current opportunity regime is not a 'buy everything' signal. The decision navigator should be used with specific household profiles. ZIP-level analysis reveals that opportunity is concentrated in 4 of 10 ZIPs for the median household. The topological buy signal is neutral overall (S_B ≈ −0.004).

Analyst conclusion The post-correction market is structurally stable but price-elevated. This creates fragmented pockets of opportunity: certain ZIP- income- DTI combinations show positive buy signals even at elevated rates, particularly in ZIPs where incomes have caught up to post-shock prices (Brookside, Patrick Henry). The opportunity is not uniform — it is topologically localized.

7. Risk Flags

8. Methodology Note

TTAS embeds property-month observations into a 12-dimensional feature space and studies the topology of sublevel sets under a tri-parameter filtration (affordability, spatial density, opportunity score). The engine computes persistent homology, Euler characteristic surfaces, persistence vineyards, and a household-specific rent-vs-buy path integral.

Data sources: FRED (mortgage rates, CPI, unemployment), Census ACS (ZIP-level income and home values), Realtor.com Research Data (listing prices, rents, inventory, days on market). When real data is unavailable, a deterministic Tulsa-calibrated synthetic generator (seed 918) is used.

Limitations: The 12-dimensional feature space is synthetic at the per-property level even when calibrated to real aggregates. Census ACS data currently returns structural headers but null metric values for Tulsa ZIPs. OSMnx spatial features are opt-in. This report is a research demonstration — not appraisal, lending, legal, financial, or investment advice.

9. Source Manifest

Source	Status	Nature	Confidence
FRED (Federal Reserve Economic Data)	active	Observed	High (0.90–0.95)
Census ACS 5-Year Estimates	active (null values — structural fetch only)	Observed	Medium-High (0.85)
Realtor.com Research Data	active — 1 CSV imported, 119 monthly rows	Observed (aggregates) / Calibrated (per-property)	Medium-High (0.70–0.85)
OSMnx Street Network + POI	opt-in — requires --use-osmnx flag	Observed (when enabled) / Synthetic (default)	Medium (0.55) when enabled; Low (0.40) when synthetic
Tulsa County Assessor / City of Tulsa Open Data	planned	Planned (Observed)	n/a