Data scientist Vs code / notebook

It really depends on the project. For quick prototypes and smaller tasks, Jupyter notebooks are great because they allow for an interactive and hands-on approach. However, when it comes to scaling the project or working on production-level code, I switch to VS Code or another IDE because it’s easier to manage large codebases, handle version control, and modularize the code. Many teams use a mix of both, starting with notebooks for exploration and then moving to VS Code for development.

Yaar, I like Jupyter notebooks because they allow for easy collaboration. You can add comments, explain your thought process, and share the results directly in the notebook. It’s especially useful when working in teams where other members can follow your work or contribute to it. Plus, the integration with libraries like Matplotlib, Seaborn, and Pandas makes it incredibly efficient for data scientists working on research and analysis. So for me, it’s more about the convenience of the notebook formats.

Earlier I used Jupyter notebook for EDA and then Spyder for code development. However I shifted to VS code since it can easily sync with code repos plus it has a lot of plug ins. Now I create an environment and install jupyter notebook in it. And open jupyter notebooks from the VS code terminal. We can install notebook plug in VS code but I don't find that user friendly. This approach allows you remain within VS code env but easily switch to Jupyter when required. Hope this helps !