IBM Data Science Professional Certificate: Tools and Platforms (Jupyter, Watson Studio, GitHub)

Tools and Platforms (Jupyter, Watson Studio, GitHub)

This domain tests the working environment: notebook execution, Watson Studio project assets, and Git/GitHub collaboration. Questions often ask why a notebook fails to reproduce or which Git command moves work from local to remote.

Core rules

• Jupyter notebooks have cells and kernels; execution order matters more than visual order.
• Before sharing, restart the kernel and run all cells; then confirm outputs match the code.
• Common magics include %matplotlib inline; shell commands can be run with !.
• Watson Studio organizes work into projects with assets, notebooks, environments, and deployment spaces; API keys and project tokens control access.
• Git has three main areas: working directory, staging area, and repository. git add stages, git commit records, and git push publishes.
• Branch, commit, open a pull request, review, and merge. Use .gitignore for data and secrets.

Worked example

A collaborator cannot reproduce your notebook. Reasoning: check the kernel and run all cells from a clean state; remove absolute paths; pin dependencies in requirements.txt; set random seeds; and confirm the README explains data access. If the notebook contains an API key, revoke the key, move it to an environment variable, and rewrite history only if necessary. Share the repository, not just the .ipynb file.

Traps

• Committing API keys, tokens, or personal data.
• Using git add . and including large datasets.
• Force-pushing over shared branches.
• Trusting saved outputs without rerunning the notebook.
• Hard-coded absolute paths that break on another machine.
• Forgetting that Watson Studio environments and credentials are project-specific.