A 'dataset' in Re:infer is made up of a user-defined group of similar sources, and will have a model associated with it that has been trained on the data within that dataset.
The model encapsulates the purpose of this dataset – i.e. what is the user trying to understand from their data?
For example, a dataset could include all sales conversations within an organisation, across multiple sources, and the model could have been trained by the user to monitor the customer experience from these conversations.
You can view a list of all of the datasets in your project on the Datasets page.
An example dataset card from the Datasets page
Please Note: You should only add multiple data sources to a dataset if they are of a similar type and share a similar intended purpose (e.g. capturing customer feedback, or multiple email inboxes that service similar requests).