View identity graphs

Last update: 2023-09-25
  • Created for:
  • Beginner

Learn how to use the identity graph viewer feature to search, explore, and filter identity graphs for validation and debugging. For more information, please visit the identity graph viewer documentation.


I’d like to explain the identity graph viewer feature. The persona we have in mind is the data engineer who’s responsible for ingesting data and maintaining data quality. The issue that these data engineers face is that they have no way to understand how identities are stitched today in intuitive manner. In other words, it’s a black box. Without this as a data engineer, you can’t ensure that the data ingested correctly, and this becomes a customer satisfaction issue as customers can’t trust the data ingested on the platform. In a worst case, they may be stitching profiles that should not be stitched and activating them. Our feature allows the user to search, explore, filter identity graphs. The benefits is that this shortens time to value and this is especially useful in scenarios where data engineers want to validate or debug the identity data after ingestion. Now let’s see this feature in action. Identity graph viewer allows the user to investigate how identities are stitched together. After entering an identity value, they’re presented with the graphical representation of the graph. You can explore this to understand how identity values are associated with one another. When clicking on these nodes in the graph, you can see that the table that corresponds to the value is also highlighted and the user is also presented with some additional information to the right that can aid in debugging. If they’re more interested in the data sources and how they are used to construct the graph, the data source tab is available for use as well. This shows essentially a timeline of how data sources were ingested and how they have manipulated the creation of the graph.

When selecting one, you can see the values that have been linked because of it as well as additional information about it.

For the case where we might be investigating an issue with a graph collapse, you could see here we’ve got two clusters that are obviously stitched together possibly an error right here. So, if we’re interested in this edge, we can click this and learn that there are three batches in this particular data set that have contributed to this edge. We know when they happened, and the batch ID, and the source. So, this would be a great starting point for debugging further. -

On this page