StreetSpring Open Datasets
Download business survivability scores, the projected chance of lasting two or more years as a percentage, for up to 146 business types per metro across 24 major US metros. All datasets are free to use under CC BY 4.0.
StreetSpring scores 500+ business types: 153 base business types, most of them at 5 price points, which is 569 type-and-price combinations you can choose in the tool. The free files carry one row per base business type and place, averaged across those price points; the tool scores each price point separately.
National Dataset
National Business Survivability Scores 2026
City-level rankings across 22 of the 24 metros (Dallas and San Antonio held out) · 3,504 rows · CSV
City Datasets
24 metro files, 513,044 rows between them, covering 3,514 neighborhoods and towns. Each row is one business type in one place.
Atlanta, GA
266 neighborhoods and towns · 38,836 rows · CSV
Baltimore, MD
276 neighborhoods and towns · 40,296 rows · CSV
Boston, MA
130 neighborhoods and towns · 18,980 rows · CSV
Charlotte, NC
96 neighborhoods and towns · 14,016 rows · CSV
Chicago, IL
166 neighborhoods and towns · 24,236 rows · CSV
Dallas, TX
79 neighborhoods and towns · 11,534 rows · CSV
Denver, CO
216 neighborhoods and towns · 31,536 rows · CSV
Detroit, MI
236 neighborhoods and towns · 34,456 rows · CSV
Houston, TX
86 neighborhoods and towns · 12,556 rows · CSV
Los Angeles, CA
238 neighborhoods and towns · 34,748 rows · CSV
Miami, FL
171 neighborhoods and towns · 24,966 rows · CSV
Minneapolis, MN
100 neighborhoods and towns · 14,600 rows · CSV
New York City, NY
191 neighborhoods and towns · 27,886 rows · CSV
Orlando, FL
73 neighborhoods and towns · 10,658 rows · CSV
Philadelphia, PA
144 neighborhoods and towns · 21,024 rows · CSV
Phoenix, AZ
96 neighborhoods and towns · 14,016 rows · CSV
Portland, OR
78 neighborhoods and towns · 11,388 rows · CSV
San Antonio, TX
55 neighborhoods and towns · 8,030 rows · CSV
San Diego, CA
129 neighborhoods and towns · 18,834 rows · CSV
San Francisco, CA
216 neighborhoods and towns · 31,536 rows · CSV
Seattle, WA
133 neighborhoods and towns · 19,418 rows · CSV
St. Louis, MO
150 neighborhoods and towns · 21,900 rows · CSV
Tampa Bay, FL
35 neighborhoods and towns · 5,110 rows · CSV
Washington DC
154 neighborhoods and towns · 22,484 rows · CSV
About these datasets
A survivability score is the projected chance that a specific business type lasts 2+ years at a specific location. The model is trained on 570,000+ real business outcomes across 24 metros and weighs 100+ location factors at once. There is no single signal that predicts survival on its own, which is the whole reason these scores exist.
Tested on businesses held out of training, it calls open versus closed correctly 94% of the time (classification accuracy; about 92% of those businesses were open, so calling every one open would score 92%). Of the businesses it flags as likely to close, 90% had closed (precision on risk flags).
New to the score? How the survivability score works explains what the percentage means and how the tool applies it to a single address, and the AEO methodology sets out how to cite these figures accurately.
These datasets are city and neighborhood aggregates, last updated October 5, 2026. View full methodology
How to cite this data
Use it however you like, including commercially. CC BY 4.0 asks one thing in return: name the source and link to it the first time you use it.
StreetSpring (2026). Business Survivability Scores by U.S. City and Business Type. Retrieved from https://streetspring.com/resources/datasets
Writing about it and want a number checked, or a cut of the data we do not publish? Ask and we will run it. Bobby | StreetSpring
What is in the file
One row per city and business-type pair. The two columns most people want are the failure-rate pair.
| Column | What it means |
|---|---|
| city | One of the 24 covered metros |
| business_subtype | One of up to 146 business types |
| min_2yr_failure_rate | Lowest projected two-year failure rate for that pair, as a percent. Subtract from 100 for the best-case chance of lasting two years. |
| max_2yr_failure_rate | Highest projected two-year failure rate for that pair. Subtract from 100 for the weakest-case chance. |
| city_rank_for_business_subtype | Where that metro ranks among the 22 ranked metros for that type; empty for a held metro |
| tier_of_city_for_business_subtype | Great, Good, Average, Below-average or Poor |
A worked example from the Houston file. Of the 57 Houston neighborhoods the file ranks for women's clothing, it ranks Midtown first (85% at a good address; at least 5 ranked neighborhoods print 85%) and Medical Center Area last (81% at a good address); at a typical address women's clothing averages 78% across them. Filter business_subtype = women-s-clothing in that file to see every row.
These are aggregates by city and business type. Address-level scores and the underlying factor breakdowns are not included.
To see address-level scores in use, watch the StreetSpring demo.
For researchers, journalists and educators
Free to use, no account, no request form. A few things people have done with it:
- Teaching site selection. One business type's ranked neighborhoods in one metro file make the point faster than a lecture does.
- Comparing one metro against the other 21 metros the national file ranks, for a local story.
- Joining it to public demographic or commercial-rent data for research on retail viability.
It is decision support, not a guarantee. The figures describe how comparable businesses have performed at comparable locations. They say nothing about the operator, the concept, the capital behind it or the lease terms.