I'm a researcher, not a developer, and I mainly use Python for simulating mathematical models and verifying experiments. Lately, I've been diving into deep learning and reinforcement learning, specifically physics-informed machine learning. My coding journey has been primarily with Jupyter notebooks and Jupyter Lab. I've heard that VSCode might be more efficient and better for managing larger projects, although my code tends to be messy and efficiency isn't my priority; I mainly need it to work so I can tweak parameters and equations as needed. However, I'm curious about how switching to VSCode could impact my work and in what ways it might benefit me. Can anyone shed light on the differences and the advantages of making this switch?
6 Answers
VSCode has a great Jupyter notebooks extension that allows you to work on Python notebooks right within VSCode, which I find much more efficient than using them in a web browser. Plus, being able to interact with your file system directly is a big plus for project management.
VSCode boasts a massive community and an extensive range of plugins that you won’t find in other tools. Jupyter might be highly targeted for researchers with some niche tools, but many folks who give VSCode a spin end up sticking with it. No rush—do what feels right for you!
From my experience coding in both, VSCode really shines due to its ability to install extensions that offer linters and syntax corrections, making your coding experience smoother. Jupyter has some of these features, but they might not be as accessible depending on your version.
If your code is under 1,000 lines and involves a couple of files, the management benefits may not be very noticeable. The argument for switching isn't super strong in that case.
As a DevOps engineer, I've noticed our data scientists gravitating towards Jupyter notebooks in VSCode, even though it can be tricky to set up with things like AWS SageMaker. Benefits like extensions and linters in VSCode are appealing, but ultimately, Jupyter notebooks work the same way regardless of where you're running them.
Hey there! Much like our research process, the best way to determine what suits you is to try them out. Give both Jupyter and VSCode a shot, and see which one you vibe with the most!
Actually, my current project is over 3k lines, and I’ll need to extend it after publishing. I want to streamline my process so I can focus on the key parts rather than getting lost in the details.