Notebook

A Notebook is a working session that lets users interact with a dataset without modifying the original data.

What is a Notebook?

When a Notebook is created from a dataset, DashAI makes a mutable copy of the original dataset. The source Dataset record is never modified. Within a Notebook, users can:

• Run Explorers (scatter plots, histograms, box plots) to visualize the data.
• Apply Converters (StandardScaler, PCA, SMOTE, etc.) to transform the dataset copy.
• Revert any converter to restore an earlier state.
• Save the modified dataset as a new Dataset record available for model training.

Database Representation

A Notebook is stored in the Notebook table, linked to the source Dataset. Each Explorer or Converter applied within the Notebook creates an Explorer or Converter record. Converters are applied sequentially via a ConverterJob.

Lifecycle

Original Dataset ──(copy)──► Notebook Dataset
                                    │
                       Apply Explorers   (read-only visualizations)
                       Apply Converters  (in-place on mutable copy)
                       Revert Converters (restore earlier state)
                                    │
                            Save ──► New Dataset (available for training)