Learn Python by Building Data Science Applications

Adding code to the equation

Now, let's say that we want to keep track of the metrics for each consecutive iteration of the code, using a specific dataset. There is no simple way to do that with a Jupyter Notebook (this could be done using a papermill package: https://github.com/nteract/papermill ), so we copied the code from the notebook and stored it in the script, predict_result.py. This takes data and writes down accuracy and model settings to the metrics.json file.

One nice feature of VS Code is that it can convert Jupyter Notebooks into code and back. Just type import jupyter in the command palette and select the notebook. It will still require some tweaks but saves a sufficient amount of time anyway.

With that (and the script itself ...