Accepted Solutions
Hey @elkinsbe! This is one of my favorite features of Visual Analytics - you absolutely can add those clusters to a table and even assign clusters to new data automatically. Visual Analytics has built-in scoring methods for its models. Models that you create in Visual Analytics can be scored directly within the report - no extra code needed. Here's how easy it is with a clustering model.
- Right-click the clustering model and select "Derive cluster ID items..."
- In the pop-up box, give the variables a name. Cluster ID will be a variable that identifies the cluster a particular row belongs to. You also get some other statistics, such as distance from the cluster centroid. This helps you identify where the observation lies compared to the cluster centroid.
- Check the data pane. You should now see your derived clusters there:
That's all there is to it! You can add these data items to any visualization you see fit in Visual Analytics. As new data comes in for this table, they'll be assigned a cluster automatically based on your model.
Whenever you model in Visual Analytics, right-click and look for "Derive predicted..." or some variation of that. That's your way to add your model predictions directly to a report. The beautiful thing about it is that a data scientist can create a model in a report, hand it off to a report designer, and the report designer doesn't need to know how the model works. They can work with the predictions exactly like any other data item.
Hey @elkinsbe! This is one of my favorite features of Visual Analytics - you absolutely can add those clusters to a table and even assign clusters to new data automatically. Visual Analytics has built-in scoring methods for its models. Models that you create in Visual Analytics can be scored directly within the report - no extra code needed. Here's how easy it is with a clustering model.
- Right-click the clustering model and select "Derive cluster ID items..."
- In the pop-up box, give the variables a name. Cluster ID will be a variable that identifies the cluster a particular row belongs to. You also get some other statistics, such as distance from the cluster centroid. This helps you identify where the observation lies compared to the cluster centroid.
- Check the data pane. You should now see your derived clusters there:
That's all there is to it! You can add these data items to any visualization you see fit in Visual Analytics. As new data comes in for this table, they'll be assigned a cluster automatically based on your model.
Whenever you model in Visual Analytics, right-click and look for "Derive predicted..." or some variation of that. That's your way to add your model predictions directly to a report. The beautiful thing about it is that a data scientist can create a model in a report, hand it off to a report designer, and the report designer doesn't need to know how the model works. They can work with the predictions exactly like any other data item.
