Onyxia is an open-source web application designed to facilitate the setup of a state-of-the-art work environment for data science.
What sets Onyxia apart is its focus on accessibility for data scientists who may not be well-versed in cloud technologies. Through a simple and engaging interface, users can specify their preferred tools (e.g., Jupyter, RStudio) and allocate computational resources (e.g., GPU, CPU, RAM), among other options. Onyxia then automates the setup process, from launching a container to configuring access to S3 data storage and managing security credentials. This allows data scientists to focus solely on their data-related tasks.
Onyxia is not the application where data scientists actually perform their work—that role is filled by tools like Jupyter, RStudio, or others. Instead, Onyxia serves as a step in their workflow for configuring their technical stack. Additionally, Onyxia features a file explorer to simplify the management of large datasets.
While Onyxia enables the effortless setup of a ready-to-use environment, it also maintains a high level of transparency regarding the actions it takes on behalf of its users. Onyxia displays the commands it runs, allowing users to gain a better understanding of the underlying processes and even attempt to run the commands themselves. This level of openness allows Onyxia to function not only as a facilitative platform but also as a learning resource. It helps demystify complex systems such as Kubernetes, Helm, Docker, S3, Vault, and OIDC for data scientists.
To utilize Onyxia, system administrators install it on a Kubernetes cluster, either on-premises or via a cloud service provider, and grant access to the web UI to their data science team.
You can find a video introducing the project here.