본문으로 건너뛰기
Analyst Path

Create your first collection and register a dataset

Create an empty collection, register a CSV as a dataset, and add the metadata that makes it findable.

8

Once you know how to find data, it's time to build your own workspace. A collection is basically a folder for one analysis project. A common rule of thumb is "one analysis project = one collection."

Create a collection

  1. From the + Create menu in the upper-right of portal, choose Collection.
  2. Pick a name. Others can find it through search, so use something that captures the project context (e.g. 2026Q3-retention-analysis).
  3. Write a one- or two-line description of the collection's purpose. This shows up as a preview when hovering in the tree and becomes a search match target.

After you click Create, the collection appears in the left tree right away. No refresh needed.

Register your first dataset

Inside the empty collection, click Add dataset and you'll see three source options:

  • CSV / JSON upload — Fastest path. Good for sample data under 10k rows.
  • External connector — Pull directly from production systems like databases, S3, or REST APIs.
  • Copy a resource from another collection — Bring an already-registered dataset into your workspace.

This lesson uses CSV upload. If you don't have a CSV handy, you can grab one from a small scenario in the dhub2-examples repository.

Once the upload finishes, the dataset is added to the collection. Click it and you'll see a preview table with inferred column types.

A one-line metadata pass

A dataset's metadata is the lowest-cost work that makes other people able to find and trust the resource.

  • Description — One line. "What data, at what point in time, at what granularity."
  • Tags — Biggest effect on search. Domain, team, and cadence are usually enough.
  • Owner — Yourself or the data owner.

Self-check

  • Does the new collection appear in the left tree without a refresh?
  • Is the uploaded CSV registered as a dataset, with the preview table rendering properly?
  • Are the description, tags, and owner each filled in with at least one line?

Next lesson

Now you'll put your first dashboard widget on top of this dataset.