Create a Workbench Instance – How to Start
Overview
This overview will be very short, because a Workbench instance can be started simply by clicking on the Workbench button in the main menu.
This feature is aimed at highly technical users such as data scientists, programmers or technical data analysts who are proficient in the Python programming language. A Jupyther Notebook is embedded in the workbench, which as open source represents a worldwide de facto standard for data science.
Noreja allows you to build directly on the graph database and perform in-depth analyses with Python directly on the knowledge graph. Scripts can be saved and extended to build extremely customized analyses.
Additional Background
Historical roots and technical core
Jupyter spun off from IPython in 2014 with a language-agnostic design and now supports dozens of languages, from Python to Julia and R. In 2015 only about 200 000 notebooks lived on GitHub; by early 2021 the number had climbed toward ten million—hard evidence of explosive uptake. The architecture strictly separates user interface and kernel through an open messaging spec, so the same notebook runs unchanged on a laptop, a cloud VM or a Kubernetes cluster. That decoupling is what makes the format so portable and durable.
Reproducibility as a scientific driver
Computational notebooks bundle code, results and narrative in one file, squarely addressing modern science’s core challenge: traceability. Studies show that notebooks help researchers share analyses transparently and make methods easier to reuse. Popular-science outlets have noticed as well—The Atlantic headlined “The Scientific Paper Is Obsolete” in 2018 and pointed straight to the Jupyter model. Nature even named Jupyter one of the ten software systems that changed science in 2021.
Community and ecosystem growth
Open governance, regular community sprints and substantial grants (Sloan, Moore, Helmsley and others) have spawned an ecosystem far beyond the core project. The VS Code Jupyter extension alone surpassed 40 million installs by mid-2022, ranking among the most-downloaded in the Marketplace. In 2017 the project received the ACM Software System Award—honours previously reserved for giants like UNIX and Java.
Industry adoption and cloud integration
Every major cloud vendor now ships notebook-centric services—Amazon SageMaker, Google Colab, Azure Notebooks and more. Netflix reports that notebooks are “the most popular tool for working with data” in-house, powering tens of thousands of daily jobs. Deep hooks into production data pipelines mean Jupyter is no longer just for exploration; it is increasingly a production-grade interface.
Why do data scientists love it?
| Motivation | Rationale |
|---|---|
| Exploratory speed | Interactive cells let you test a hypothesis in seconds rather than minutes or hours. |
| Narrative communication | Markdown, plots and widgets embed context right next to the code. |
| Language & environment freedom | One notebook can mix Python, R and SQL—just switch kernels. |
| Rich ecosystem | Libraries like pandas, Matplotlib and scikit-learn are optimised for notebook workflows. |
| Reproducibility out of the box | The full analysis path—from raw data to figure—remains version-controllable. |
Surveys back this up: JetBrains’ 2023 Developer Ecosystem survey found that at least 35 % of data professionals use Jupyter notebooks, and over 40 % of those spend more than a fifth of their work time inside them. In 2022 roughly 40 % named Jupyter their preferred editor for data-analytics tasks, and nearly 70 % for machine-learning modelling. Python’s meteoric rise amplifies the effect—the language’s vast ecosystem and gentle learning curve feed directly into Jupyter, whose “native” kernel is IPython.