The ValiFit Metrics Catalog

Up-front and annual MIP.

15-year rate.

Drives all loan calculations.

State-level median for "below-median" filter.

State income tax for relocation comparisons.

County FHA cap; drives /api/affordability.

Source for health composite county-level rollups.

County-level median household income.

Share of sales closing above asking. Direct heat indicator.

Median price of homes that closed during the period

Year-over-year % change in median sale price

Share of new listings going pending within 14 days. Demand signal.

Median of last sale price across town parcels. Strong national coverage where parcel_details has sale data: NJ/CO/CT/FL/MA/MD (~60-80% of parcels).

Months to clear inventory at current sales pace. <4 = sellers market, >6 = buyers market.

Median time from list to pending. Low = hot market.

Median sale price per square foot — best size-normalized price comparison

Total homes for sale at end of period

Median price of homes newly listed during the period

YoY change in months-of-supply. Rising = inventory accumulating = buyer leverage.

YoY change in price-drop frequency. Rising = cooling market = buyer leverage.

% of active listings that had a price reduction during the period — buyer-leverage signal

Supply-side flow indicator

Forward-looking demand indicator

Most-recent town-level median sale price; falls back to state aggregate if town volume too low

CBSA name + metro code containing the city

Start of the data period (typically a month or 28-day rolling)

city / county / metro / state / national

HUD FMR vs user budget for ACTUAL bedroom count from intake. Catalog merged with id=56 May 7.

Property tax rate. Cascade: tax_rates_v2 (NJ-only) → municipality_metrics → parcel_details (where parcel-derived rate computable). Catalog merged with id=54.

avg_sale_to_list < 100 = buyers negotiating below ask = leverage. > 100 = sellers market.

Median monthly HOA across active listings in town. Affects total monthly carry beyond P&I.

Tax bill on a median-priced home in this town

HUD-published Fair Market Rent for 2-bedroom unit in this metro area

Median monthly gross rent (Census ACS B25064)

Effective rate as % of market value (NJ-style, normalized for assessment ratio)

Distinct routes serving stops within 0.5mi of town centroid. Higher = more options.

Distinct transit modes (bus/rail/subway/ferry) within 0.5mi. Diversity = resilience.

Census ACS B19013 — used in budget composite + price-to-income ratio

Total resident population (ACS B01001 1-year for places ≥65k, 5-year otherwise)

NCES F-33 exp_total / enrollment. Catalog merged with id=64.

Composite ROI: outcomes per dollar of per-pupil spending

Elementary English Language Arts proficiency rate (state-administered assessments)

Land-area % within Special Flood Hazard Area (Zone A/AE/V/VE)

low / moderate / high / extreme — derived from flood_zone_pct

1 = high (>4 pCi/L expected), 2 = moderate, 3 = low

Distinct active NPIs / 1k residents. Catalog merged with id=66 May 7.

Share of providers whose primary_specialty matches HRSA primary-care taxonomy codes.

Health-insurance coverage gap; signals access barriers

Count of NPI-registered active healthcare providers practicing in town

1-5 star CMS quality rating averaged across hospitals in 10mi radius

Capacity ratio — drives hospital_capacity composite

CMS-certified inpatient facilities within service radius

0-26 — higher = greater shortage severity

True iff town is designated Health Professional Shortage Area (HPSA)

Years — primary health-outcome metric

Years of potential life lost before age 75 per 100,000 population

Cumulative price change over 5 years; equity-build proxy

Census ACS rental vacancy rate for the metro; lowers investor confidence at high values

Year-over-year median rent change

NRHP-listed sites within 1mi of town centroid. Indicator of preservation overlay restrictions / character.

Median effective_year_built across town's parcels. Effective accounts for major renovations.

Median living_sqft from parcel_details. National coverage in NY/FL/MA/MD/CT (62-86%). Falls back to active_listings.sqft.

Cascades parcel_details (10+ states market_value) → parcels → active_listings

Share of active listings flagged new_construction. Indicates supply growth + neighborhood evolution.

Raw murder count (FHA-safe — used in safety_roi denominator only, not bias-correlated)

Year of the most recent UCR/NIBRS submission used for crime metrics

Direct link to FBI Crime Data Explorer for this jurisdiction

NFPA 1710/1720 standard service radius

ROI denominator for safety_roi_officers_crime composite

Burglary + larceny + auto theft per 1,000 residents/year

Murders + rapes + robberies + aggravated assaults per 1,000 residents/year

Total students enrolled (K-12)

SES proxy: (free + reduced) / enrollment. Used in school equity analysis.

Format: "PK-5" / "K-8" / "9-12" — derived from grade_low + grade_high

elementary / middle / high / k12 / other

Full-time-equivalent teacher count (drives student_teacher_ratio)

True iff school is charter (publicly funded, independently operated)

Students eligible for free lunch (federal poverty guideline ≤130%)

Regular / Special education / Vocational / Alternative

Students eligible for reduced-price lunch (130-185% federal poverty)

12 = large city, 21 = mid-size suburb, 41 = rural fringe, etc. (NCES locale codes)

True iff school is fully online (no physical campus)

Sum K-5 enrollment columns

FHA-FLAGGED: descriptive only. Used in Phase 9.I bias audit; NEVER enters scoring (CI test enforces).

7-digit federal district identifier

12-digit federal school identifier; canonical join key for school-level data

FHA-FLAGGED: per-race counts. Used ONLY in Phase 9.I quarterly bias audit + persisted as numerical correlation matrix; raw counts NEVER influence scoring or display. CI test (Phase 6.A) enforces.

TIGER EDGE shapefile GEOID for spatial joins

15-digit GEOID for census block; used for tract-level demographic enrichment

Net operating income / property value; primary cash-flow yield metric

Annual pre-tax cash flow / cash invested; primary investor leverage metric

(annual rent / price) * 100

Annual rent − operating expenses (tax, HOA, insurance estimate, maintenance reserve)

Price / annual gross rent; lower = better deal (3-7 fair, <3 great, >10 weak)

True iff num_units >= 2 (small mf) — strategy filter

The bilateral match% from Atlas Invariant 1 — same % shown to both sides of a dyad

Hours until match expires per Atlas Invariant 2 (24h for buyer↔agent high-fit)

JSONB array of why this match scored well (rendered as bullet list in JSON-LD + UI)

excellent / very good / good / average / fair / poor

Deed recording date

Deeded sale price from county recorder; primary signal for appreciation

Sum of land + improvement; basis for property tax

Improvements (buildings) portion of assessment

Land-only portion of assessment (excludes improvements)

Assessor's adjusted year reflecting renovations — vs raw year_built

Assessor's estimated full market value (some jurisdictions equate to total_value)

True if any pool (in-ground, above-ground, or community)

Year of last major renovation; falls back to effective_year_built from assessor

Most recent annual tax bill from assessor records; alias property.annual_tax for source variant

True iff cooling system is central AC (not window units)

Assessor's quality grade (low / fair / average / good / very good / excellent)

finished / unfinished / walkout / crawl / slab

Integer count; bool source (has_fireplace) coerced to 0/1+

Second-most-recent sale; pair with prior_sale_date for appreciation signal

Value attributed to special features (pool, deck, finished basement)

Bedrooms + bathrooms + living rooms + kitchens (assessor count)

Yearly tax bill on the parcel.

Tax-assessor valuation.

NCES ID of the elementary school for the address.

NCES ID of the high school.

NCES ID of the middle school.

Listed bed count; matches user "I need N+ bedrooms".

Colonial / ranch / cape / split-level / contemporary / etc.

Standardized class (frame / masonry / steel-frame / etc.)

Free-form description of construction type from assessor

central / window / split / none

Per-county unique parcel identifier (APN, PIN, etc.). Format varies by jurisdiction.

0-100 — completeness of property record (drives confidence intervals on derived scores)

Book number of deed registration — useful for title searches

Page number of deed registration

Computed from list_price + current rate + standard 20% down

Used for cap rate / cash flow on Investor composite

FEMA designation for the parcel polygon.

Full bath count.

Covered parking count.

Half-bath count.

forced air / radiant / heat pump / baseboard / etc.

monthly / quarterly / annual / one-time

Standardized land-use designation (single-family, multi-family, commercial, etc.)

When ValiFit last spatial-joined and enriched this parcel

Date of most recent recorded sale.

Most recent recorded sale.

Heated/finished living area.

Total lot footprint.

True if newly built.

Total baths from assessor; full + half combined (vs property.full_baths/half_baths)

Building count on the parcel (main house + ADU + outbuildings)

Signals in-law / accessory units when > 1

Numeric stories from assessor; alias property.stories for listing-source variant

Multi-family unit count (1=SFH, 2-4=small mf, 5+=apartment building)

Per-county OPRA/sunshine-law disclosure. May be person, LLC, trust, or government.

Town / city / borough — joined to municipalities_canonical for scoring

Raw assessor address + ValiFit-normalized variant

SFH/condo/townhouse/etc. enum.

JSONB blob of raw assessor fields (state-specific schemas) for forensic / debugging access

NCES district name; canonical for join with school_performance

public / septic / private / none

Building stories.

Recorded subdivision name from plat map

Year the annual_tax value applies to

Total parking (garage + driveway + carport + street-assigned)

public / well / shared / cistern

Original construction year.

Local zoning designation (R-1, R-2, mixed-use, etc.)

Full marketing description — primary input for listing vibe-vector embeddings

Realtor.com's automated value estimate — used for estimate_to_ask_ratio

When status flipped (active→pending→sold/withdrawn) — drives DOM bucket

Tax-assessed value from county records — denormalized onto listings for speed

Array of canonical tags pulled from public_remarks (doorman, pre-war, turn-key, etc.)

Comma-separated additional photo URLs (max ~25 per listing)

When same listing appears in multiple feeds, one row is canonical

Days since first listed in current cycle.

Ratio of estimate to ask; >1 means underpriced.

Monthly HOA dues.

When ValiFit last pulled this listing from source

Date originally listed.

Current asking price.

Authoritative listing-agent attribution.

MLS description; embedded into listing_embeddings for vibe match.

active/pending/contingent/withdrawn enum.

sale / rent / lease / exchange / etc.

Local MLS ID — distinct from listing_key (which is realtor.com's id)

Source MLS the listing originated from (NJMLS, Bright, MRED, etc.)

Date listing went pending.

Number of MLS photos (proxy for marketing effort).

List price ÷ living sqft.

Realtor.com canonical listing identifier — primary join key

Match of property to declared strategy (buy-hold / BRRRR / flip / 1031 / Section-8)

Occupied units / total units across portfolio

Days unit sits vacant between leases — primary efficiency metric

minimal / standard / strict / institutional — affects match-rate with renters

Voucher acceptance (note: source-of-income discrimination protections vary by state)

Self-managed vs PM-managed; informs matching with PM services

How complete the user's profile is — gates pool eligibility per Atlas Invariant 3 (agent ≥80%, buyer ≥60%)

Renter-declared days until needed move-in; drives lease-urgency tier

Annual income / annual rent; standard landlord screening = >40x monthly

Renter has pets; filters listings without pet-policy match

Self-reported, optional. Affects approval-confidence calculations.

Affects approval at landlord screening

6 / 12 / 24 / month-to-month preference

JSONB array (in-unit-laundry, parking, doorman, gym, dishwasher, etc.)

studio / 1br / 2br / 3br / townhouse / sfh-rental

In NYC/Boston-style markets — willingness to pay broker fee

Used by /api/affordability multi-loan compare.

Combined with FHA/Conv minimums to set max purchase.

Routes school weighting to elem/middle/high tier.

Geocoded destinations + frequency for haversine commute composite.

Per-composite weight in the user-controlled match.

Unlocks FTHB DPA programs in education layer.

deal-breaker | strict | default | flexible | totally-flex per priority.

Affects bed/bath need + DTI denominator.

Triggers school-composite weighting.

cash_flow | appreciation | flip | brrr.

Months. Affects rental match radius.

Brewery/hiking/dog park etc; weights livability composite.

CFPB QM 43% back-end DTI cap.

Replaces mortgage math; drives rental match.

Yard/pet-friendly building need.

Monthly rent / price target.

False = no new leads will be routed to this agent

Lower is better; correlates with conversion. Tracked by lead_assignments timing

True iff agent's license has been verified against state board record

unverified / pending / verified / suspended

5-star average from past clients; outcome-validated against close rate

Free-form bio used for vibe-vector embedding + agent profile pages

True iff agent has authorized ValiFit to sync their IDX feed

NAR designations (CRS, GRI, ABR, etc.)

JSONB array (FTHB, luxury, investor, relocation, multi-family, etc.)

JSONB array of ISO 639-1 codes

Zips/towns the agent actively services.

Composite reputation score; visible to buyers.

Mean days from intro to close.

Per-state real estate broker license number for the agent's brokerage

Count of in-flight lead_assignments — used to enforce capacity

Timestamp of most recent successful pull from agent's IDX feed

Pct of intros where agent reached buyer within SLA.

Salesperson license number per state real-estate licensing authority

Multiple Listing Service the agent reports through (e.g. NJMLS, Bright)

Numeric NAR member ID — confirms REALTOR® designation

JSONB array of county FIPS codes the agent serves

JSONB array of municipality slugs the agent serves

FTHB / luxury / investor / relocation etc.

free / lite / agent_pro / brokerage / office

Travels across the n-sided graph: AQS for agents, response-rate for buyers, occupancy for landlords

Cross-portfolio vacancy; lower = better operator

Tours that result in signed lease; primary leasing-agent quota metric

Net operating income divided by unit count

1-5 average tenant rating

apartment / sfh / condo / mixed-use / commercial