Creating an experiment

Click New Experiment in the navigation bar. Pick your storage provider at the top of the form and give the experiment a Title. The title becomes the name of the Drive folder, Dataverse dataset, or Zenodo deposition that DataPipe creates for you, so choose something you'll recognize there. You can rename the experiment later from the pencil icon beside its title, and its ID stays the same so your code keeps working, but the name in your storage account is set once, at creation. Rename it there yourself if you want the two to match.

The rest of the form depends on the provider. DataPipe only asks for what the provider needs, and each field explains what the provider does with it:

• Google Drive: nothing else is required. If you want the data somewhere specific, click Choose Drive folder and pick a parent folder. Otherwise DataPipe creates the folder in My Drive/DataPipe.
• Dataverse: the collection alias, which is the short name from your collection's URL, plus the author name, contact email, and description that Dataverse requires for every dataset. It won't create one without them. The contact email starts out as the one on your DataPipe account. Change it if you want a different address published on the dataset. Subject is optional and defaults to Social Sciences.
• Zenodo: the author name and a description for the deposition. Zenodo would let you create a draft without these, but it won't let you publish one, so DataPipe asks now rather than leaving you to find out at the end of your study. Affiliation is optional.

Click Create. DataPipe makes the Drive folder, Dataverse dataset, or Zenodo deposition for you and opens the experiment dashboard, which links straight to it.

A new experiment isn't accepting data yet. It starts with data collection off, validation on, and no session limit, so nothing can be submitted until you turn on data collection from the dashboard.

You pick the provider once per experiment, and every file from that experiment lands there. Choosing a provider

The dashboard is where you find the experiment ID your code needs, watch sessions arrive, and change settings later. Open it from My Experiments.

• The heading shows the experiment's name, whether it's accepting data, and how many sessions have completed.
• The details block below it holds the experiment ID, the value your code sends with every request, and a link that opens the Drive folder, Dataverse dataset, or Zenodo deposition in your own storage account.
• Data collection, Validation, Metadata, and Finalize are the settings for this experiment. You can change any of them while a study is running, except finalizing, which is permanent.
• Integration code has ready-to-paste snippets with your experiment ID already filled in.
• The queued files panel appears only when DataPipe has accepted a submission that hasn't reached your storage provider yet. Most of the time you won't see it.

The completed-session count is the number of submissions DataPipe accepted, including any still queued for your provider. It's not a count of the files in your storage, and it doesn't include rejected submissions or base64 file uploads.

A queued upload is retried automatically, and you can download it from the dashboard in the meantime. When an upload fails

The Data collection section has four switches. Each takes effect immediately, and none of them touches data you've already collected.

• Accept new data: while this is on, your experiment accepts submissions from participants. Turn it off and every submission is rejected with DATA_COLLECTION_NOT_ACTIVE. Leave it off until you're ready to recruit.
• Accept base64 file uploads: a separate switch, needed only if your experiment sends binary files such as audio, video, or images. Media and binary files
• Assign conditions in sequence: gives each participant the next condition number in order. Set how many conditions you have when you turn it on. Condition assignment
• Stop after a set number of sessions: refuses submissions once the count reaches the limit you set, so a study can't overshoot its recruitment target. Session limits

Two more settings sit below them, each with its own page: Validation, which is on for every new experiment and checks submissions before they reach your provider, and Metadata, which describes your data in a standard format as it arrives.

Every switch you turn on widens the path into your storage provider, so turn each one off again when collection ends. Security posture