
Storefront risk, measured.
Restaurant, payment and public data, combined at the level of one store.
Every store, in three layers.
Public records, card sales and POS operations meet at one address.
Three layers. One dataset.
Matched by business name and address. Public layer live today; card and POS layers from 2027.
Test decisions on accurate data.
Checked on years the model never saw, in two cities.
How it works
Three kinds of data, joined at one store.
Openings, closures and the street around a store
How sales move, with merchant consent
What happens inside the store
How the score is calculated
From public records to a risk band.
What we look at
Business age and industry, staff count, nearby competition, foot traffic and transit access, and recent closures in the area.
How it learns
A statistical model trained on storefront openings from 1998 to 2021, then scored on 2022–2023, years it never saw.
What you get
Owners see break-even sales and area comparisons. Lenders see a closure-risk band for each store.
Model weights and data-cleaning rules are not published.
One engine, two screens
For owners and for lenders.
Where to open, and what you keep
Compare areas and see the sales you need to break even.
When, how much, and on what terms
Risk bands checked on years the model never saw.
Evidence
Tested on years the model never saw.
A storefront closure is not a loan default. These figures support decisions next to a credit file; they do not replace it.
Demo
The demo is available by invitation.
For lenders, partners and program reviewers.
Or email [email protected]