Exploring collected data and model versions.

The data drawer makes it easy to explore the data that was used to train each iteration of a users machine learning model. They could easily toggle beteween different models to test how they performed with additional taining data.