> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.umnai.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.umnai.com/_mcp/server.

# Data

Data workflows in the platform turn source files into datasets that can be used for model training. The flow has two main stages:

* **Ingestion** validates source data and creates a datasource.
* **Onboarding** turns that datasource into a dataset ready for training.

## How data moves through the platform

Organise the source files you want to use for ingestion.

Give the platform read-only access to the storage location that contains your source data.

Describe how each source column should be interpreted, including field types, designations, and encoding options.

Run ingestion to validate the source data and create or update a datasource.

Select features, define the target, and prepare the datasource for model training.