Bayesnote is a frictionless integrated notebook environment for data scientists and data engineers. It provides a user interface to build dashboards and deploy machine learning models right from a notebook. It also supports the operation of notebooks by a workflow system, Noteflow. It manages servers, libraries, and containers for development and production.

Bayesnote v0.0.1 demo

The motivation of building Bayesnote is explained in this Bayesnote: Redefine Notebook.

Bayesnote is consisted of

1. Unified Notebook Backend. It is designed to reuse the computation engine by integrating with existing notebooks, like Jupyter notebook and Zeppelin notebook, and to be integrated by other apps like Apache Airflow by exposing notebook operations, e.g. run/interrupt, as REST APIs.
2. Noteflow. It is a workflow system built for notebooks rather than functions. The dependency of notebooks, called Noteflow, is specified in YAML and triggered by cron or events. It is written in Go rather than Python to make it easier for developers/contributors to maintain and easier for users to deploy.
3. Dashboard component. It enables data scientists to process data in a notebook with Python, SQL, R and Spark, and build dashboards right from the notebook.
4. Machine Learning component(under development). It provides one-click deployment to production right from the notebook.
5. Container component. It automates common tasks of managing containers for notebook users for environment isolation and deploy to production.

Bayesnote also introduces experimental features into the notebook itself:

1. Multiple language support. Write Python, SQL, R, and Spark in one notebook by selecting a language for each cell of the notebook.
2. Variable sharing. Variables in different languages in one notebook would be automatically shared across cells.

We introduce a few examples to demonstrate the progress of development of Bayesnote:

1. Write Python, SQL, R, and Spark in one notebook. Users can select a language for each cell and get results printed.

#data-engineering #data-science #workflow #jupyter-notebook #data analysis