Get Started

Explainability with Dataiku

Understand model outputs, increase trust and eliminate bias.


Predictive Model Explainability

Dataiku provides critical capabilities for explainable AI, including reports on feature importance, partial dependence plots, subpopulation analysis, and individual prediction explanations.

Together, these techniques can help explain how a model makes decisions and enable data scientists and key stakeholders to understand the factors influencing model predictions.


Disparate Impact and Subpopulation Analysis

Models built with biased data will produce bias predictions.

Disparate impact analysis in Dataiku measures whether a sensitive group receives a positive outcome at a rate close to that of the advantaged group. Dataiku also includes sub-population analysis that allows users to see the results by group. Both analyses help to find groups of people who may be unfairly or differently treated by the model.

With this information in hand, data scientists can produce models that deliver more responsible and equitable outcomes.


What-If Analysis

Dataiku what-if analysis allows data scientists and analysts to check different input scenarios and publish the what-if analysis for business users.

With what-if analysis accessible to business users, they can build trust in predictive models as they see the results they generate in common scenarios and test new scenarios.


Individual Prediction Explanations and Reason Codes

Dataiku provides the individual prediction explanation report to explore the data and results at a row level.

During scoring, in both batch and real-time, prediction explanations can be returned as part of the response, which fulfills the need to have reason codes in regulated industries and provides additional information for analysis.


Automatic Model Documentation

Dataiku Model Document Generator automatically creates necessary model documentation based on a standardized template.

With standardized and automated documentation, organizations maintain consistent records of AI projects for regulatory compliance, and data scientists can get back to work building more projects and creating more value.

Go Further

White Box vs Black Box Models

How to balance interpretability and accuracy, stemming from the difference between black-box and white-box models.

Read on the blog

Get a Demo

Watch our end-to-end demo to discover the platform.

On-Demand Dataiku Demo

Read the Responsible AI In Practice Series

What happens when AI systems make a mistake? Why does that matter?

Read on the blog

Discover how Dataiku enables Analytics Leaders

Beyond silos, explore how Dataiku helps teams with quick wins and long-term transformation.


Get Started with Dataiku

Start an online hosted trial, download the free edition

Get Started