What are the differences between the two solutions?

• Dataiku focuses on facilitating collaboration between data scientists, data analysts and business users, offering a data analysis and machine learning platform that supports the end-to-end development of AI projects. It is designed to make AI projects accessible to all user profiles, offering visual tools for data preparation, modeling and model deployment.
• Databricks, on the other hand, is a unified platform for big data and machine learning, designed to facilitate the processing of large quantities of data using Spark clusters. It is particularly renowned for its data processing performance and its ability to perform complex large-scale analysis and modeling tasks.
• The two platforms can be complementary in a data ecosystem. Dataiku can be used for its ease of use and ability to enable different users to collaborate on analytics and AI projects, while Databricks can be chosen for its high-performance data processing and optimized environment for big data and advanced machine learning.

Pricing for Databricks

Databricks is billed on a pay-per-use basis, meaning that users only pay when they use the Databricks platform.
Databricks uses an internal consumption metric, the Databricks Units, or DBU. The number of DBUs required to perform operations varies according to location (America, Europe, Asia, ...), cloud provider (AWS, GCP, Azure) and machine selected. The more powerful a machine, the greater the number of DBUs required to use it. If the user uses several machines, this multiplies the number of DBUs required to perform the operations. Each operation (SQL queries, DLT advance compute, etc.) has a price per DBU.

Example: For a compute job at 0.15€/DBU, active 8 hours a day, on 5 instances (or virtual machine) costing 1DBU per instance, the Databricks cost will be : 5 instances x 1 DBU/hour (machine price) x 8×30 (compute 8 hours a day, every day) x 0.15 DBU (job price) = €180/month.

Pricing for Dataiku?

Dataiku offers a variety of plans to suit different team sizes and needs, from a free offer to more advanced enterprise options. Here's an overview:

1. Free Edition: intended for permanent installation for up to 3 users, enabling access to open source files or databases, to be installed on your own infrastructure.
2. Discover: Designed for small teams of up to 5 users, including more than 20 connectors and the ability to process in-memory or in-database (Spark) with limited automation.
3. Business: Suitable for medium-sized teams of up to 20 users, offering unlimited and elastic computations with Kubernetes, full automation and advanced security, but with limited deployment.
4. Enterprise: Provides scalable automation and governance, including all connectors, full deployment capabilities, isolation framework, and unlimited instance and resource governance.