On this page
Distribution Reference
Every int/float column carries a distribution in its distribution_params. Pick the shape that matches the real-world quantity; pass the listed parameters. All distributions honour min/max clamping and decimals.
{"type": "integer", "distribution": "poisson", "lambda": 4, "min": 0}Continuous#
| Distribution | Params | Use for |
|---|---|---|
normal | mean, std | Symmetric quantities: age, height, test scores. |
lognormal | mu, sigma | Right-skewed money: price, salary, order amount. |
uniform | min, max | Flat ranges: latitude, a 1–5 rating. |
exponential | scale | Wait times, inter-arrival gaps. |
beta | a, b (scaled to min/max) | Bounded proportions, scores in a fixed band. |
gamma | shape, scale | Positive skewed durations, insurance claim sizes. |
Discrete#
| Distribution | Params | Use for |
|---|---|---|
poisson | lambda (alias lam) | Counts per interval: items per order, calls per hour. |
binomial | n, p | Successes out of n trials: conversions, defects. |
Heavy-tailed (power-law)#
For "a few get most, most get very few" — views, followers, wealth, file sizes:
| Distribution | Params | Notes |
|---|---|---|
zipf | a | Discrete power-law; larger a ⇒ steeper tail. |
pareto / power_law | alpha, scale | Continuous heavy tail. |
{"type": "integer", "distribution": "zipf", "a": 2.0, "min": 1} # view counts
{"type": "integer", "distribution": "binomial", "n": 10, "p": 0.3} # successesConditional & correlated shapes#
Distributions describe a single column's marginal. To make columns relate:
- One column switches another's distribution →
depends_on. - Two numeric columns move together → correlations.
- A quantity changes over time → outcome curves (magnitude) or rate curves (proportion).
Parameter aliases#
A couple of names are accepted both ways so hand-written and tool-generated schemas both work:
- Poisson rate:
lambdaorlam. - Zipf shape:
a; Pareto shape:alpha.
In the studio#
The column Inspector auto-suggests a distribution from the column name (e.g. views → power-law, price → log-normal) as a one-click preset, and the Engine params panel exposes the full distribution picker with the relevant shape fields.
Prefer generating data visually?
Misata Studio lets you turn natural language prompts into relational schemas with exact outcome invariants, live preview tables, and one-click database seeding without writing code.