Glossary
| Word | What it means |
|---|---|
| Story | What the data must show, in your words: the numbers, the window and the events ("MRR from $50k to $200k over two years, churn spiking in Q3"). |
| Stated number | A number in the story that Misata holds exactly: a monthly total, a balance, a rate, a share or an average. |
| Held | The data shows the stated number, read back in SQL. "Not held" means it could not be, and the dataset says why. |
| Window | The months the data covers. The last month is the one a dashboard opens on. |
| Event | A change for a while: a spike, a dip, a campaign, an outage. "5% in Q3 2025" is an event on a 2% rate. |
| Built-in shape | A business Misata knows in depth (SaaS, e-commerce, banking and payments, health systems and more), with its tables, mechanics and standard metrics. |
| Plan | What Studio will build, shown before any rows exist: tables, sizes, stated numbers, window and events. |
| Certificate | The record of what was checked on the finished data: each stated number, each rule, each key, and anything open. Shareable as a public link. |
| Verified | Every stated number met, every declared rule held, and no critical or major finding open. |
| Realism notes | Observations about the data (seasons, hours, concentration), with their SQL. They describe; they do not change the verdict. |
| Answer key | Each metric's SQL and its value in every month of the window, as answer_key.json and a CSV. |
| Metric (KPI) | A number the business reads: MRR, logo churn, billed revenue, win rate, or one you stated. |
| Misreading | A usual wrong way to compute a metric (churn over every account, not those at risk), measured on the same rows so a wrong answer can be named. |
| Identity | An equation the data keeps, read in SQL: billed revenue equals MRR; an order's total equals its lines. |
| Checker | Runs a query, or reads numbers, against the answer key. Names a right one; says what a wrong one computes and how to fix it. |
| Questions | Plain-language questions drawn from the answer key, each with its answer, its SQL and the wrong answers its misreadings give. |
| Seed | The number that fixes the random draws. The same request and seed make the same data, byte for byte. |
| Variation | The same story with a new seed: different rows, the same stated numbers. |
| Revision | A change to a ready dataset (in words, a story edit, a curve or a control). It makes a new version and checks it; the first is never touched. |
| Curve | A monthly total you draw; each month lands on it. |
| Control | A named dial on a dataset (how strongly a cause acts, staff at peak, planted problems), changed on the Adjust tab. |
| Planted problems | Data-quality issues put in on purpose (duplicates, late records, typos), each listed, for testing cleaning and monitoring. |
| Blueprint | The engine's own design language: tables, columns, rules, processes and stated numbers, in JSON. What a design becomes before it is built. |
| Backlot demo | A pipeline (dbt or Postgres) seeded so its own output tells a story, with a room, a reset and an agent endpoint. |
| Demo room | A Backlot link anyone can open with no account: the story, what the pipeline measured, and a read-only SQL console. |
| Instructions | Rules every dataset on your account keeps to ("company names fictional"), read as requirements. |
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.