Within the Re:infer platform, data is structured and stored in a hierarchical manner, which comprises of 3 main components - data sources, datasets and organisations. Access to each of these is controlled by strict permissions.


Data Sources

These are collections of raw unlabelled communications data, e.g. all emails from a shared mailbox or a collection of NPS survey responses (see here for more detail).


These are comprised of 1 - 20 similar data sources and the 'model' that you create when you train Re:infer to understand the data within those sources (see here for more detail).


A permissioned storage area within the platform. Each dataset and data source belongs to a specific organisation, which is designated when they are created (see here for more detail).


These are per-user and specific to each organisations that a user belongs to. They can provide access to sensitive data and, depending on the permission, allow users to perform a range of different actions in the platform (see here for more detail).


The below illustration helps to understand the relationship between these components and permissions.

In this example, all of the data sources are associated with Organisation A, whilst there are datasets associated with both Organisation A and Organisation B.

If a user wanted to access datasets in Organisation A (i.e. 1, 2 or 3), they would require viewing permissions for Organisation A only.

But if a user wanted to access datasets in Organisation B (i.e. 4, 5 or 6), they would require viewing permissions for both Organisations A and B, because the data sources are all located in Organisation A.




Example data structure within Re:infer

Next: Create a Source in the GUI