Skip to Content
3. Create a Dataset

Create a Dataset

A dataset is a single transformation run: raw files mapped into a schema and taken through review and approval. Datasets always belong to a schema, so the target structure is fixed before you upload anything.

Steps

Open a schema

From your folder, open the schema you want to transform data into.

Go to the Datasets tab

The Datasets tab lists every dataset created from this schema and its current state.

Start a new dataset

Click New Dataset to launch the transformation wizard.

The transformation wizard

A dataset moves through four steps, shown in the step indicator at the top:

  1. Upload Raw Data — bring in your Excel/CSV files and choose worksheets.
  2. Processing — AI transforms the raw data and maps it to your schema.
  3. Review — inspect the mapped rows and approve or reject.
  4. Export — write the finished data out (and merge into the main schema).

Dataset states

As it progresses, a dataset carries one of these states:

StateMeaning
DraftCreated but not yet submitted.
Pending ApprovalSubmitted and waiting on a reviewer.
ApprovedApproved and ready to merge.
RejectedSent back — needs changes.
CompletedMerged into the main schema table.

Long-running work (transformation, exports, merges) runs as background jobs, so you can leave the page and check back — the dataset’s state reflects progress.

Next step

The first wizard step is uploading your data: Upload Excel Files →

Last updated on