Surface machine learning predictions in Tableau Next visualizations and dashboards with the Predict() function defined in the semantic model. Compare historical data and predicted outcomes side by side in a visualization to identify trends, risks, and opportunities. Access prediction scores, drivers, and recommendations right where you visualize data and take action faster. Use prediction fields in calculated fields, filters, and aggregations to refine and extend your analyses.
Where: This change applies to Tableau Next in Enterprise, Performance, and Unlimited editions.
Why: For example, after analysts add prediction fields to a sales dashboard, sales reps can compare historical performance with predicted outcomes. When they identify a low likelihood-to-close score, they can investigate the prediction drivers and take steps to revive the opportunity.
How: To create a prediction field, create a Predict entity in the Semantic Model Builder. Select an existing machine learning model and bind the model inputs to semantic model fields. Choose additional insights, such as top drivers and recommendations, and save the definition.