Skip to main content

Hi guys,

I’ve been trying to set up Data Graph. What I’d want to do with it is (I think) pretty straightforward. I’d like to use Linked Events to pull traits as they are known in Unify to populate those on events.

I’m having trouble configuring the Data Graph so that it will actually allow me to pull these traits in mapping to any destination.

At first I assumed that pointing to the Profile Sync dataset in our DWH would be enough for Segment to enable this. After all, Segment creates these tables with default schema’s so I figured it should work out of the box. I learned that this is not the case and that you’d need to set Enrichment first before you can select traits to populate on events.

I set up an Enrichment that points to the user_traits materialized view, but that doesn’t work as there is no user_id in there to match on a property of any event.

Currenlty the data graph looks like this:

data_graph {
    version =  "v1.0.0"

    profile {
      profile_folder = "PROJECT_NAME.segment_profiles_sync"
      type = "segment:materialized"
    }
    entity "account-entity" {
     name = "Profile Traits Unify"
     table_ref = "PROJECT_NAME.segment_profiles_sync.user_traits"
     primary_key = "id"
     enrichment_enabled = true
}
}

Does anyone have a template for a straightforward Data Graph setup that can pull traits from a profile as they are synced with Profile Sync?

Thank you!

Hi ​@Wouter Hol! A couple follow up questions to make sure we have a good understanding of what you’re trying to do: 

  • What’s the ultimate use case you’re trying to achieve?
  • What’s destination are you sending this data to?
  • What event(s) are you trying to map and what data points does the event(s) need to be enriched with?

This table is by trait and not in a format that linked audiences would use so a suggestion is to create a new table within snowflake or your dwh where you have a single column with user id, email etc. and any other traits you want to generate and pull events from.

https://segment.com/docs/unify/data-graph/linked-events/#step-5-enrich-events-with-entities

https://segment.com/docs/unify/profiles-sync/sample-queries/#pull-a-complete-list-of-your-customers-along-with-their-merges-external-identifiers-or-traits

Hey ​@Wouter Hol 

Our data graph looks something like this (i’ve had to strip some out to make a simplified version):

data_graph {
  version = "v1.0.0"
  profile {
    type           = "segment:unmaterialized"
    profile_folder = "XXXXXXX"

    relationship "users-to-profiles" {
      name           = "user-to-profiles"
      related_entity = "user"
      external_id {
        type     = "user_id"
        join_key = "USER_ID"
      }

      relationship "accounts" {
        name           = "user-to-accounts"
        related_entity = "account"
        join_on        = "account.USER_ID=user.USER_ID"
        relationship "products-to-accounts" {
          name           = "products-to-accounts"
          related_entity = "product"
          join_on        = "product.ACCOUNT_ID=account.ACCOUNT_ID"


        }
      }
    }

  }

  entity "user" {
    name        = "User"
    table_ref   = "XXXXXX.XXXX_USER"
    primary_key = "USER_ID"
    // For Linked Events - optional and defaults to false
    enrichment_enabled = false
  }

  entity "account" {
    name        = "Account"
    table_ref   = "XXXXXXX.XXXX_ACCOUNT"
    primary_key = "ACCOUNT_ID"
    // For Linked Events - optional and defaults to false
    enrichment_enabled = false
  }

  entity "product" {
    name        = "Products"
    table_ref   = "XXXXXXX.XXXX_PRODUCT"
    primary_key = "PRODUCT_ID"
    // For Linked Events - optional and defaults to false
    enrichment_enabled = false
  }

}


This ends up with an entity relationship diagram like this:
 

Entity relationship diagram

Now when you create a Linked Audience you can traverse up and down these entities and retrieve data from other columns related to the profile.

Eg. I can create a Linked Audience of all the Profiles that have exist in the User table, that are related to a Account that have a Product, where Product table has column “product_sku” = “123”

When mapping this audience to a destination, you can then also use data in these related entities to pass to the destination. Eg. the SKU that the customer has can be populated in the mapping to be sent to your conversion destinations.

Hope that makes sense?

Reply


Terms & Conditions