PROJECT · LAUNCHED NOVEMBER 2026
Scarcity
Independent research on whether the supply of trained people matches workforce demand. It starts with jobs where shortages are often reported. For each one, it can find a shortage, enough supply, or not enough evidence to say.
PROFESSIONS
24
In behavioral health, schools and speech-language, nursing, and other health jobs.
PAGES
1,200+
State-by-profession pages, all built from the same code and the same rules.
COST TO USE
Free
The research and data are free to use and download, with credit, under CC BY 4.0.

WHY
Why I’m building it
I bought scarcity.co in 2012. I wanted an easier way to see which jobs were most needed and most valued, and to spot trends early. That information was spread across government files, licensing boards, and reports that cost a lot of money.
I studied economics in college and business analytics after that. This data is public, and it should be free and easy to use.
THE PROBLEM
Why this is hard
The numbers that bear on a shortage sit in separate places, and they often define a job differently.
- A count of people finishing training, a count of new licenses, and a count of people working in the job answer different questions.
- Schools report the same kind of program under different codes, and some programs lead to a different license than their name suggests.
- Projected openings are not unfilled jobs, and a projection only fits the people it actually counts.
Most of the work is deciding which numbers belong together, checking them against each other, and saying what they can and can’t show.
THE METHOD
How it works
The rule behind it
Claim only what the evidence supports. When it doesn’t, say we don’t know yet.
Each profession page counts the people who finish training, checks that count against other sources, compares it with projected need, and gives one status label. The sources are public: federal education data, the national provider registry, federal workforce projections, and licensing boards.
- Labels come from written rules, applied the same way to every profession. No one sets a label by hand.
- Every key number is tagged as observed, a projection, or our estimate.
- Every source file is recorded with its release, the date it was pulled, and a checksum, so anyone can check a number against the file it came from.
- When a number changes, the page shows the old and new numbers.
- The data refreshes every month, and every change is checked automatically before it goes live.
No school, company, or group can pay to be listed, ranked, or featured. Every school is counted by the same rules. The research and data are free to use.
How a page gets built
Four steps, run the same way for every profession
Pull public data
Federal education files, the national provider registry, federal job projections, and licensing boards.
Check the counts
Compare each count with other sources, and keep only programs that lead to the license.
Apply written rules
Each profession gets one label from rules that are the same for every job. No one sets it by hand.
Publish with sources
Every page shows its sources, tags each number, and keeps the old numbers when they change.
Labels at launch
Not every job that people call short shows a shortage
- 7SignalThe evidence points to a shortage.
- 8WatchOne source points to a shortage, or the sources disagree.
- 3No shortage shownThe evidence doesn’t show a shortage.
- 6Not enough evidenceThe evidence isn’t strong enough to say.
24 professions, as of the November 16, 2026 release. Labels can change with each monthly data update. A fifth label, Structural, is for a shortage that has held three years or more. No profession has it yet. The rules
THE BUILD
How I build it
I build Scarcity with Claude, an AI model made by Anthropic. It helps write the code that builds the site, drafts page text, and helps read source documents. The numbers come from public data files and are computed by that code. Labels come from written rules, not from an AI’s judgment. I review every change before it goes live.
Before launch, I had the site reviewed against the strongest criticisms I could find. Among them: that someone who works in higher education built it to sell programs, that the free site is a front for paid tools, and that counting graduates isn’t the same as counting supply. Each one led to changes. The site now explains who runs it, the conflict of interest, how its data is used, and what graduate counts can’t tell you. (About Scarcity)
STATUS
Where it stands
Scarcity launched in November 2026. Next up: outside reviewers for the labeling rules, permission from a few data owners, and more professions. The full plan is public on its What’s next page.
BUILD NOTES
What I’m learning
-
I downloaded a list I wasn’t allowed to republish
A site with no robots.txt still had terms that banned automated copying. I took the data down, then wrote rules for what we use and when we ask first.
-
My site looked finished. AI tools saw almost none of it.
My pages looked done in a browser. With scripts off, a full page showed about 4,300 characters and no title. Here’s the fix, and how to check your own site.