Model once. Query everywhere.
Four building blocks
Datasets snap together like building blocks: Functions, Code Files and Models combine into bigger Models — and every block stays individually executable.
Functions
Parameterized queries for business users — the everyday building block.
Code Files
R or Python scripts executed in the sandbox's secure runtimes. The twist: their output is itself a dataset. Your Python script becomes a refreshable table in Excel.
Models
The special entity: a Model composes Functions, Code Files and even other Models into one virtual view that returns tabular data — nested, reusable, governed.
Value Lists
Built on a Model but returning exactly one column, made to be bound to a parameter — the result: governed dropdowns in every client.
Location → System Line → System → Virtual Dataset → any client
Typed parameters
Every Virtual Dataset can define any number of typed parameters — strings, numbers, dates — each optionally backed by a Value List. That is what makes datasets feel like products instead of queries: in Excel the sidebar wizard renders parameters as form fields with dropdowns, in the Excel add-in and the R/Python packages they are passed as parameters, in Power BI they are filter values.
One dataset, every stage
A System Line groups sibling Systems — dev → QA → prod, or an old system next to its successor — with one System as the line's default. Datasets bind to the line, not the box: they execute against the default, can be pinned to a dedicated System where needed, and at runtime any client or the Test Bench can switch the System — strictly within the line, so flexibility comes with guardrails.
Your landscape has stages — dev, QA, prod. Your reports shouldn't need a copy per stage.
See your own data live in 30 minutes.
Start with a live demo — we'll show your use case in Excel, Power BI, R, or through an AI assistant. Then take it further with a free trial in your own isolated tenant.