Dataiku Adds New Visual Features and White Box AI to Dataiku 6

Dataiku Adds New Visual Features and White Box AI to Dataiku 6

Source: Dataiku on YouTube

Dataiku has released the latest version of its AI and machine learning platform, Dataiku 6, according to a press release on the company’s website. Version 6 is highlighted by the ability for users to spin up and manage Kubernetes clusters from inside the Dataiku platform. The product also features a suite of capabilities for building sustainable AI systems. There’s a new plugin store that enables customers to extend the platform, as well as improved subpopulation analysis for better model performance and avoiding model bias.

Our Buyer’s Guide for Analytics and Business Intelligence Platforms helps you evaluate the best solution for your use case and features profiles of the leading providers, as well as a category overview of the marketplace.

Dataiku offers an advanced analytics solution that allows organizations to create their own data tools. The company’s flagship product features a team-based user interface for both data analysts and data scientists. Dataiku’s unified framework for development and deployment provides immediate access to all the features needed to design data tools from scratch. Users can then apply machine learning and data science techniques to build and deploy predictive data flows.

Users can spin up and manage Kubernetes on AWS, Azure or Google Cloud Platform from inside Dataiku. According to the release, this means that “non-admin users can now quickly spin up Kubernetes clusters for optimized execution of Spark or in-memory jobs.” Administrators can also isolate and manage compute power so teams get what they need to run analysis and utilize AI. Snowflake users will gain faster runtimes with optimized sync with WASB and native execution of Spark jobs. Version 6 hastens the process of executing multi-step SQL data pipelines as well, allowing for optimized compute and storage.

Dataiku 6 touts partial dependence plots and subpopulation analysis that lets users deep-dive into key aspects of model behavior. Subpopulation analysis provides the ability to weed out unintended model biases to create transparency and fair AI deployment. Partial dependence plots help users understand complex models visually by surfacing relationships between a feature and target. Cross-team collaboration add-ons include improved IDE integrations for RStudio, VS Code, SublimeText and PyCharm let coders work in any environment. In sum, the update makes the process of working with external data visualization tools more efficient.

In a media statement about the news, Dataiku CEO Florian Douetteau said: “Dataiku 6 enables enterprises to do just that by offering more features for white box AI, collaboration, efficiency, and elastic resource management to allow businesses’ AI to evolve along with the technology.”

Learn more about Dataiku 6 in the company’s blog, featuring Sunny Porinju.

Timothy King
Follow Tim