Extending and personalizing your Discourse Graph

This page covers ways to extend and personalize Discourse Graph beyond the default setup.

Stored relations

If you want to use the current relation workflow, start with the Stored relations guide.

Pattern relation legacy

Discourse Graph used to rely much more heavily on query-based pattern relations. The videos below are older examples of that legacy workflow.

Note:

They use the legacy query-based pattern-relations workflow
They show an older UI that does not match the current product

Some users extended the grammar quite extensively. For example, here is a video from a user who extended the grammar to cover structured international law research:

https://www.youtube.com/watch?v=rST7cKMO_Ds

Others also extended the grammar to integrate results from their own experiments and contextualize each new result against earlier research.

For example, here is a demo of creating a “consistent with” relation pattern that recognizes when two pieces of evidence both support the same thing:

https://www.loom.com/share/cb9e526a98764e95a459a6db2b66e46a

If you want to explore or understand that legacy workflow, it will be useful to learn more about the extension grammar.