23.12 Release Notes

 

Metric Drivers and Insights

The 23.12 release of Max brings a new analysis type known as Metric Drivers. Metric Drivers analyzes the performance of a chosen metric and describes the impact on that performance from other, related, metrics. These supporting metrics are outlined by the user in the metric hierarchy.

 

Back to Top

Metric Hierarchy Management

In order to run a Metric Drivers analysis, you must first build out a metric hierarchy. This "tree" of sorts will inform Max which metrics have an impact on the main metric in question.

To setup a metric hierarchy:

  1. Navigate to the desired dataset by clicking on the datasets icon.
  2. Select your dataset from the list.
  3. Toggle to Properties.
  4. Click on Metric Hierarchy.
  5. Click the plus icon or Create a Node to begin adding metrics to your hierarchy.
  6. Use the metric dropdown to select a metric from your dataset. If you wish for the metric to be called something other than its name in the hierarchy, change its label.
  7. Continue adding metrics to your tree until your hierarchy is complete.
  8. Click on a metric and use the trash can icon to delete it if necessary.
  9. Deleting the node will remove it from the hierarchy and move any applicable children up in the tree. Deleting the node with its children will remove that entire branch from your hierarchy.

Back to Top

Metric Drivers from Follow-Ups

 

With the 23.12 release, you can now ask a Metric Drivers question directly from the suggested follow-up questions in Max's response. The suggested follow-ups will be generated from the top 3 drivers that Max identified in the original analysis. To run a follow-up question, simply click on the suggested query.

Back to Top

 

Driver Analysis Follow-Ups from Table

Driver analysis questions now offer automatic follow-up suggestions in the query results. To run a follow-up question, simply click on the suggested question and Max will run the analysis.

 

Back to Top

 

Bulk Dataset Management

In the 23.12 release, owners of datasets can now copy datasets to create a group of datasets. This group of datasets makes it easier on owners to perform bulk management actions across the applicable datasets. Owners can copy concept configurations (for both domain and calculated metrics) to other datasets in the same group.

Once the dataset has been copied, it can be shared to the same users/group as the original dataset.

To copy a dataset:

  1. Click the datasets icon.
  2. Choose the dataset you wish to make a copy of.
  3. Click the three dot menu next to the dataset's name.
  4. Select Duplicate Dataset.
  5. From the modal, review the new group's name and the resulting dataset's name and click Duplicate Dataset.
  6. A toast will appear letting you know when the new dataset and group have been created.
  7. You will now see a group icon above the Properties menu. Click the icon to review the datasets in the same group as this dataset.
  8. Here you can detach this dataset from the group if desired.
  9. Now, you can make necessary changes to metrics in your dataset. Once your changes are complete, click the three dot menu next to the metric name.
  10. Select Copy Metric to Group.
  11. In the modal, review the properties of the metric that will be copied to the group. Changes to the metric's properties will be highlighted here for confirmation.
  12. Check the boxes of the datasets in the group that you wish to copy the changes to.
  13. Select Confirm when you are ready to save the changes.

Some things to note that occur when a dataset is duplicated:

  • The new dataset will be named with the format: “{Dataset Name} (copy)”
  • The new dataset keeps the original description
  • Entire matching domain config included calc metrics is ported
  • Any existing dataset metric hierarchy is ported
  • Any existing filters are ported
  • All statistics are copied
  • Generates embeddings
  • The new dataset (and original if not yet done) is attached to a group that is named based on the original dataset

Back to Top

Composite Keys

In the 23.12 release users can now manually create joins that require multiple fields, or composite keys.

If Max does not recognize a match when adding a join to a dataset, then Max will provide a link for adding a composite key join.

Follow this link to create your composite key.

At the sample view of your dataset, confirm the details of your composite key and click, Infer Join.

Here you can provide keys for both tables that you wish to join to your dataset. Once the keys have been entered, click Validate Join.

You'll see green (ideal), yellow (workable), or red (invalid) validation outputs based on the results of your composite key validation. If your join is green or yellow you can proceed with adding this data to your dataset.

If your join appears red, you'll be prompted to discard the join.

 

 

Back to Top

Updated

Was this article helpful?