Creating a dataset

You can find thousands of free, open data resources on data.world to work with (check out our guide to finding data), but quite often you’re going to bring in your own data to use.

When adding data to data.world, you’ll typically want to create a dataset. A dataset is simply a repository of data including data files and associated metadata, documentation, scripts, and any other supporting assets that should be stored alongside the data. Datasets can be created manually, which is what we’ll walk through here, and we also have automated options if working with larger datastores (contact us for more details).

To create a dataset, click on the + Add dropdown on the right side of the header bar and select New dataset:

After you've opened the new dataset you will:

• Add a title (up to 60 characters)
• Select an owner (if prompted) - if your dataset is for an organization, we recommend creating it under the team account to keep your organization's work within a single library.
• Choose the visibility of your dataset: 
  • Private - only accessible to you,
  • Organization - allows your team to view the project
  • Open - available to the data.world community

If you’re unsure which permissions to choose, we recommend starting private and adding contributors and increasing visibility as you go.

Next, either drag and drop one or more files into the add data box, select the Add databutton for additional source options, or save your dataset and add data at a later time.

There are multiple ways to add data and connect your data sources:

• Upload from your computer, or select cloud storage services (Box, Dropbox, Google Drive)
• Pull directly from URL or API (see the article Adding files from a URL for details)
• Connect your data via an integration (check out our super connectors if you don't see your datastore listed)

To upload files from your computer either drag and drop the file(s) from your hard drive to the Add data window or select the Add data button for more options:

So that data owners can fully document their data, data.world supports all file types. Use a script to clean the raw data? Upload it so others can see and build off your work. See What file types can I upload? for more information about how data.world handles different file formats, and when you're ready, select the Create dataset button at the bottom of the page to continue.