hello! i’m currently looking into elementary to present it to the rest of the team and hopefully integrate it into our stack to monitor data quality and be able to detect anomalies in an efficient way. I have successfully added the elementary package in dbt and it is running without issues, as I can see all of the tables the elementary package creates after executing

dbt run

. However I’m trying to just do a random test on a random column in a model, but i’m running into the following message:

Not enough data to calculate anomaly score.

(in the

test_results_description

column in table

elementary_test_results

) I have configured the test for my column as follows:

- name: my_column
        tests:
          - elementary.column_anomalies:
              timestamp_column: my_date
              anomaly_sensitivity: 6 ## forcing anomalies to show up
              days_back: 15

Is the error referring to elementary not having enough historical data to compare against in the past? Or does the test look at the results of the clusters it created by grouping over the timestamp column? I just want to be able to execute the tests and throw some results so I can present this to the rest of the BI team before we proceed with the integration 😄 Let me know if you need any other info from my side to troubleshoot - and many thanks in advance for your help!

Hey @Carlos Lopez! Really glad to hear you’re enjoying elementary Do you have data from 15 days ago? Based on the message seems like we can’t find data from the time in the

days_back

and that’s why we can’t calculate the anomaly If you’d like, and this is only for a showcase you can reduce this number and try again, or you can wait the 15 days 🙂 All that if you indeed say you don’t have data that far back in the table (if you do, we’ll continue the investigation)

thank you so much for the quick response @Noa Kurman I really appreciate it 🙏 this does indeed make sense as i think my dev data might not be up to date - i assume then that the parameter looks at days back as of today rather than the max date of the table, right? which makes total sense but I hadn’t thought of my data being stale. Let me check and come back to you