Lab · Free tool · v1.0.0
Build your own simulated population
Turn a question into an inspectable population model. Filter 39.5 million Californians from public Census microdata, weigh two kinds of scenario against the people you selected, draw attribute-only personas, and export everything with its assumptions attached.
- Define the question. One sentence. It travels with every export so the assumptions stay attached to the decision.
- Filter the population. Age, personal income band, education, employment, sex and race. Every count is weighted, so it reads as people in California, not rows in a file.
- Compare scenarios. A verified universal-basic-income policy model, and a reach-and-adoption estimate for a product, price or budget decision where every assumption is a visible input.
- Draw personas. Records sampled from your selection in proportion to their survey weight, described by attributes only. Synthetic composites drawn from Census microdata. Not real people.
- Export. A self-contained interactive HTML report, a CSV of the personas and a JSON of every input, built in your browser.
How to read the numbers: observed comes straight from a Census record · calculated is weighted arithmetic on those records · simulated is the UBI model's output · assumed is something you typed · estimate is arithmetic on assumptions and not a forecast.
Statewide baseline
Default population, no filters. Computed at build time from 12,000 Census records, each representing 3,293 Californians.
- observed Records
- 12,000
- calculated People represented
- 39,512,222
- calculated Share of California
- 100%
- calculated Median personal income
- $16,566
- calculated Mean personal income
- $38,857
- calculated Mean age
- 38.2
simulated Default UBI scenario on this baseline — $1,000/month to adults 18+, 30% flat tax on personal income: program cost $367.7B, tax revenue $460.6B, net gainers 18.34M, poverty proxy 45.8% → 32.9%. Model output on a weighted Census sample. Illustrative, not a fiscal forecast; poverty uses a personal-income proxy, not the official household measure.
Workbench
Loads 12,000 records (670 KB) into your browser on first view. Nothing you enter leaves the page. Assumption-driven estimate, not a forecast applies to every reach output.
01 Define the question
02 Filter the population
03 Compare scenarios
A · UBI policy model simulated
The verified California UBI model, applied to the filtered population. Parameters are assumptions; outputs are model results, not fiscal forecasts.
Model outputs appear when the data loads.
B · Reach and adoption estimate
For a product, price or budget decision. Every input is an assumption you make; the outputs are arithmetic on the weighted population. Assumption-driven estimate, not a forecast.
Assumption-driven estimate, not a forecast. No behavioral model, no calibration, no uncertainty range. The population is real and weighted; the rates are yours.
04 Draw personas
Synthetic composites drawn from Census microdata. Not real people. Each card is one record from your filtered selection, drawn with probability proportional to its survey weight, described by the attributes on the record and nothing else.
Persona cards appear when the data loads.
05 Export
Built in your browser from the current state; nothing is uploaded. Each file carries the question, filters, assumptions, results and the same labels and disclaimers shown here.
Mitchell Miller