Making data science a team sport
We’re spending a lot of time interviewing data scientists around the world to understand how they work and interact with their peers and other stakeholders. What have we found? Today, data science is an individual sport. As a data scientist, you have to choose your own tools and work independently. With luck, you might produce a meaningful insight, but anything you learn stays with you — it’s self-contained.
To be a data scientist today, you need to master many different skills:
- Statistics, machine learning, and optimization
- Communication and storytelling
- Programming, sometimes in multiple languages: Python, R, Scala, SQL, and more
- Big data and cloud computing
- Business and domain knowledge
- Data visualization
It is nearly impossible to find someone with all this knowledge. And with data science rapidly evolving, there are new cool tools, open source libraries, and techniques appearing every week. It’s hard to stay current, and data scientists can feel overwhelmed.
To battle these challenges head-on, IBM is investing a lot to make Data Science Experience a truly collaborative environment — a place where data science becomes a team sport. Today, we are very excited to announce the next set of features that will help to bring this collaboration to the next level:
In a Jupyter notebook, you can now have real-time discussions with peers working in or watching your notebooks. Comments are a great way to insert notes for collaborators or other viewers to see. You can add, edit and remove comments while editing or viewing a notebook.
Version control for notebooks
Now you can save versions of a Jupyter notebook. You can see a history of changes to the notebook and make previous versions current. For now, you can keep up to 10 versions of a notebook, and we’re working to extend it even further.
Now you can share the notebooks and data sets on Twitter and Linkedin directly from the Data Science Experience UI.
Like your favorite community cards
We have new community cards every day in Data Science Experience. They bring you articles, tutorials, sample notebooks, and open data sets. Now you can like them, and we will soon add sorting options in order to show the most popular items from the Data Science Community. Start liking your favorite stuff! With your input, we’ll learn your favorite topics and improve the community experience.
Dashboard view of Jupyter notebooks
You can now share an entire notebook with collaborators or share only the output graphics in a dashboard view. You can also hide sensitive information such as access credentials.
See here an example of the same Jupyter notebooks with:
Hide sensitive code cells
If your notebook includes code cells with sensitive data, such as credentials for data sources, you can hide those code cells from anyone you share your notebook with. Project collaborators with viewer permissions will also not see hidden cells. Project collaborators with admin and editor permissions will still be able to see hidden cells.
We hope you enjoy all these new features!