Dataset Versioning gives you governance and traceability over your data: every change to your dataset is captured, so you can understand what was modified and by whom, reliably reproduce results across Risk/Demand/Rate projects, and safely collaborate without losing track of the dataset state used for modeling and reporting.
Modifications that create a new dataset version
You can modify your dataset through a range of supported actions. Each sequence of modifications is automatically saved and generates a new version of the dataset.
The following actions are tracked:
Modify levels of a variable
Create formulas
Edit formulas
Change variable usage (Categorical ↔ Ordinal)
Cap variables
Rename variables
Remove variables
Export from a Rate project
Apply filters
Reuse an existing format
⚠️ Legacy datasets (created before July 2025 when versioning was introduced) will not have their past changes listed in the DB History. However, any new modifications performed after versioning was enabled will be recorded.
Dataset History
All actions performed on a datasets are stored in the Dataset History, accessible by clicking the History icon at the top of the Data Module, next to the dataset name:
The list of dataset versions — including those created by you or other users — will appear on the right side of the screen.
Each entry corresponds to a saved version of the dataset. The first entry from the bottom shows how the dataset was created. If it was created from cloning or filtering there is a link to the original database at the version the action was done from.
From the DB History panel, you can:
Clone a version to create a new editable copy of a previous dataset (Note that this is the way to edit older versions)
Name a version to easily identify it
Edit/Delete name if it has it
When you click on an older version (other than the most recent one), it will be displayed in read-only mode: within this mode you can explore variables and formulas definitions.
Dataset versioning across different modules
The versioning system is integrated throughout the platform: In the Risk, Demand, and Rate project menus, Akur8’s interface displays the “Information” tab where you can spot the dataset and the dataset version used in the project. It will show the version name if it has it or it will show the datetime point when the version was finalized:
If the model uses the latest dataset version, this will be explicitly indicated (like in the image above).
Otherwise an icon will appear, like below:
Clicking on the dataset version name in these modules opens the DB History view directly in the Data Module.