Signal Bank

(signal, 0..1) — google_maps + county permits + UCC + state license. Trailing-indicator distress score 6-18 months ahead of public bankruptcy/credit-bureau hits. Highest-ROI composite in the doc. ~50K entities expected (post UCC ingest).

(tag) — storeleads + google_maps absence + Census ACS. Premium DTC brands with no physical storefront. Used by PE / brand-acquisition firms.

(tag) — storeleads + google_maps + state license + WARN. Declining online but rising B&M reviews — the B&M strength masks ecommerce decline.

(tag) — city permits + business license + utility hookup. Zip-level composite (NOT the existing

Multi-NPI provider entities at the same canonical address with diverging state medical licenses AND IRS 990 revenue rising — the healthcare M&A buyers' target list.

Composite 0..1 civic-engagement score from FEC contributions, state SOS officer roles, IRS 990 board seats, and (when available) LD-2 federal lobbying disclosures. LD-2 is un-ingested today; the score degrades to a 3-source formula.

Owner-officer running >=2 commercial properties via >=2 LLCs with UCC equipment financing — small-fund CRE arbitrage pattern. UCC is un-ingested; emitter uses stricter property/LLC thresholds in its absence to avoid false positives.

Workforce-count divergence (>=3x ratio) across DOL H1B, EEO-1, and IRS 990. EEO-1 is not ingested today; the composite still fires on the DOL H1B vs IRS 990 2-source divergence — see the per-composite docstring for the EEO-1 dependency.

Operator group rapidly incorporating entities across >=3 states with recent SEC filings (8-K / Form D / 13D/G) and/or UCC filings against subsidiaries - the classic roll-up M&A signature before the public announcement. UCC is un-ingested; emitter falls back to a stricter (>=4 sta

B2B-heavy operator confirmed across multiple sources (IRS 990 / FMCSA / SEC / LinkedIn) with substantial scale BUT near-zero google_maps presence and no storeleads (consumer DTC). The contrarian moat: promoting the cohort everyone else penalises as low-data-richness.

Entity with at least one foreign-capital indicator -- SEC 13D/G non-US reporting person, state-SOS foreign officer, IRS 990 foreign-grant disclosure, FinCEN BOI (deferred) -- AND active US operations (Maps). KYC / compliance lens primary.

Owner-occupied commercial property with HMDA origination 2022-2024 at >=6% rate AND a distress signal (declining Maps rating, tax-delinquency, UCC) AND no active bankruptcy -- the receptive refi cohort. UCC / tax-delinquency sources are queued; emitter uses a stricter rating-drif

Franchise consolidation

Stealth expansion

Succession risk

Distressed acquisition target

Multi jurisdictional growth

Seasonal revenue cycle

Regulatory capture pattern

Cross state franchise arbitrage

Domestic supply chain resilient

Talent war signal

Architect firm consolidation

Pre construction buildup

Multi site envelope program

Tenant relocation signal

Design firm specification chain

Design firm geographic expansion

Premium finish pipeline

Anchor tenant buildout

Multi brand developer program

Corporate relocation cluster

Scaled count of DISTINCT sources confirming this entity (0..1).

Entity is confirmed by 3+ distinct sources.

In both irs_990 and sba_ppp — non-profit that took a PPP loan.

In nppes AND (google_maps OR storeleads) — provider with retail presence.

In storeleads AND google_maps — omni-channel small business.

In fdic AND nppes — anomalous: bank with healthcare NPI.

Present in any federal source AND has active phone/website.

Count of distinct federal-regulator sources, scaled to 0..1.

industry:<sector> tag from NAICS prefix (per-row Python).

size:<band> tag from employee_count (per-row Python).

Meta-signal: cross_source_count >=3 AND phone AND website AND employee_count >=10.

PPP recipient that still shows operating signals: cross_source_count >= 2 AND phone AND website.

PPP recipient with NO post-2024 / recent-activity source signals — likely dormant or closed.

Took federal money (PPP / 7a / HUD permits) AND has courtlistener history — compliance-monitor cohort.

Surviving entity at an address where >=50% of co-tenants are inactive or remerged — geographic distress cluster.

Composite 0..1 signal: employee_count >= 50, year_established >= 2018, cross_source_count >= 2, took federal money. Scales by employee_count quartile.

Storeleads (online) AND google_maps (B&M) AND (snap_retailers OR category:retail). The 'modern retail' combo.

NPPES + IRS 990 + >=2 distinct addresses across source-link payloads — hospital system or multi-site provider organisation.

Tag pair: bank_branch_active when FDIC entity is_active=true and has google_maps; bank_branch_dormant otherwise. Per-row Python deriver.

Entity at an address with >=1 HUD building permit in the last 5 years AND has a commercial NAICS prefix.

Composite 0..1 signal: 10 <= employee_count <= 100, age 5-25y, cross_source_count >= 1, has phone+website. PE acquisition target.

Composite 0..1 signal: 1 <= employee_count <= 25, took PPP/7a, recent business-license renewal, no recent court issues.

Composite 0..1 priority: court-defendant + federal regulator sources + recent inspection failures.

count of OTHER active canonical entities at the same (state, zip5, street_canonical). High counts are the classic mail-drop / registered-agent / shell-LLC pattern. ~24M coverage; the headline single-item ROI item of the plan bundle.

for each PERSON, how many distinct entities they're an officer/principal/agent on; broadcast back to every entity they touch as MAX(score).

count of OTHER active canonical entities sharing the same normalized E.164 phone.

count of distinct US states a given (normalized name) entity appears in across SOS sources.

passthrough on canonical_entities.dba_names array length. Cheap but currently not queryable as a signal.

count of UCC-1 filings naming the entity as debtor in the last 18 months. We don't have a UCC-to-entity FK so the join is via normalized debtor name + state with trigram matching to tolerate the punctuation drift between CO's debtor file and SOS registrations.

(score_0_1) Reviews in the trailing --window-days (default 30) divided by the reviews in the immediately-prior --window-days, then mapped to [0, 1] with 0.5 = flat, 1.0 = doubled-or-better. An entity with

(scalar mapped to [-1, 1]) Slope of the per-30-day review count over three contiguous 30-day buckets covering the last 90 days. Fitted via the closed-form OLS estimator for three equispaced points; output is normalised by the bucket mean so the same slope on a

(scalar mapped to [-1, 1]) mean_rating_last_90d - mean_rating_90_to_180_days_ago, divided by 5.0 (the rating scale) and clipped to [-1, 1]. None is emitted when either window has < 3 ratings.

(scalar) (current - prior_yoy) / prior_yoy where

(count, exposed as a raw float) Count of new license-issuance events in the last --window-days (default 30) in the entity's zip5. Computed once per zip5 and broadcast to every entity in that zip — same pattern as compute_hotzone_signals.py. None is emitted whe

B&M prospecting / outbound sales

Brick-and-mortar prospecting lens. Optimised for sales teams building call lists from physically-present businesses.

Analysts / BD / regulators / investors

General-intelligence dossier lens. Broadest scope; surfaces events, signals, provenance, and the explanation graph.

Healthcare BD / payer ops / medical-device sales

NPPES-centric provider lens, sorted by re-licensure recency. Surfaces taxonomy + IRS 990 healthcare enrichment fields.

Public affairs / lobbyist / federal-contracting

Every entity with a federal regulatory footprint, ranked by how many distinct federal sources touch it.

Foundation researcher / charity rating / planned giving

IRS 990-centric lens for tax-exempt organisations. Surfaces endowment / income / NTEE classification.

Real estate developer / commercial broker

Commercial-development opportunity lens. Surfaces entities with active building permits, commercial-zoned construction activity, or explicit development-opportunity tags.

Private equity firm / distressed-asset investor

Distressed-asset prospecting for PE buyers. Combines pandemic-closure risk, address-cluster decline, federal funding plus active litigation.

MCA underwriter / alt-finance broker

Merchant cash-advance prospecting. Targets active small businesses with a physical footprint and a credible revenue proxy in the underwritable band.

In-house compliance / RegTech / KYC operations

Compliance / risk monitoring. Surfaces entities with high regulatory density, anomalous cross-source presence, or active litigation flags.

Territory reps selling SaaS/payments/equipment to owner-operators

Owner-operator B&M businesses with exactly one location, an active storefront, and a reachable point of contact. Negatively weights chain / franchise membership.

Franchise-development BD, equipment-leasing reps, multi-location SaaS sales

Operators running >= 2 locations under a shared officer or chain affiliation. Sorted by location-count signal.

HQ-level brand sales (insurance, IT, supply, national accounts)

Chain-HQ-level entities: every member of a tagged chain plus the public-company-chain anchor. Sorted by chain-growth velocity so accelerating brands surface first.

Distressed-asset acquirers, contrarian operators, turnaround consultants

B&M operators showing review-velocity decay paired with cluster-decline tagging. Designed for buyers hunting distressed-but-still-open targets.

Brand-expansion site selection, retail BD, CRE leasing reps

Entities in ZIPs flagged as new-business hot zones with anchor-tenant adjacency and uncrowded NAICS-category density.

CRE brokers, retail-pad site consultants, co-tenancy negotiators

High-rating, high-review-volume entities whose adjacency draws traffic. Targets CRE roles negotiating co-tenancy or pad-site deals.

Bar/restaurant suppliers, alcohol distributors, late-night DSPs

B&M operators with material after-21:00 activity, weighted by rating and review volume. Alcohol-license sources boosted.

Seasonal-supply chain reps, staffing agencies, equipment leasing

Operators with category, hours, or review-distribution signatures of seasonality. Built for reps timing outreach to the pre-season ramp.

Outbound SDRs at vertical SaaS — Toast (restaurant tech), Square (POS), Gusto (payroll), Lightspeed (inventory)

Pre-launch business detection lens. Surfaces entities 30-90 days from operational launch (SOS formation + permits + license filings) that are NOT yet on google_maps. Target persona is the outbound SDR who wants the lead before every other vendor pounces.

Brand partnerships, premium-product placement, tourism boards

Single-location standouts with 4.3+ rating and 200+ reviews — the long tail of cult-favorite B&Ms that aren't chain-owned.

Community-bank underwriters / KYC-ops / EDD analysts pricing operator risk against the ultimate beneficial owner, not the applying LLC

Per-operator (person, not entity) risk index for community-bank underwriting and KYC EDD. Supplements compliance_monitor by weighting the operator-network diamond (Sprint 3) most heavily; surfaces every entity touched by a critical or elevated operator.