- Use Anaconda Python (Instead of standalone Python) on windows for developing python programs.
- Never install too many (or any) additional packages in the base environment.
- Create Different conda environments for different projects, e.g., one for machine learning, one for web scrapping, one for real-time hardware simulation etc.
conda create --name machine_learning python=3.7
(Optional: Specify Python version)conda activate machine_learning
conda install pandas numpy scipy matplotlib sciit-learn
(etc.)
- Install ipykernel in all environments for adding them to kernel selection menu in Jupyter UI.
conda install ipykernel
python -m ipykernel install --user --name machine_learning --display-name="Machine Learning with Python 3.7"