V-FINDER V2

A free map of PPP loan risk built from public data. It ranks all 11.47 million PPP loans by traits they share with publicly charged fraud cases, so reviewers can decide where to look first.

See a place? Question it.

Business name or street address

Typos are OK. Statistical risk indicators, not accusations.

Open full version Open Lite (phones) Methodology One-pager (PDF)

90-second demo: county map, risk profile, loan search, county audit.

County risk map, Cyber theme — County risk map (Cyber theme)

Spider chart comparing two counties — Spider chart: county risk profile

Loan detail with red flags for a publicly charged case — Loan detail with plain-English flags (a publicly charged case)

Lite version on a phone — Lite on a phone

What it does

County risk maps and spider-chart profiles
Loan search by name or loan number, with a plain-English list of red flags
County audits against SAM exclusions and OFAC
Caught Cases: how charged loans score
Gap Closer: what non-public data would catch the rest

Data (all public)

SBA PPP loan file (11.47M loans), SAM.gov exclusions, OFAC SDN, FEC, IRS exempt orgs, SBA 7(a)/504, USDA, Restaurant Revitalization Fund, Census ACS, and some state business registries.

How well it works

On 125 held-out charged loans, the application score catches 21% while flagging 0.9% of all loans, 33% at 2% and 38% at 4%. Adding forgiveness status raises that to 44%, 47% and 53%.

It misses most small sole-proprietor, stolen-identity and fake-document fraud, and most flagged loans are legitimate. Known gaps.

Built by Pat Callahan, Nashua, NH · @python0o · python0o@mail.grokbot.com · v-finder.info

Report an error or request a correction: email python0o@mail.grokbot.com. Corrections are reviewed and fixed or removed promptly.