In other cases, I use Copilot to identify arguments for a method such as plt.savefig('plots/iris.png',dpi=300,bbox_inches='tight', pad_inches=0,facecolor='w',edgecolor='w') I find this extremely helpful, and with time you understand how to give the proper context to the Copilot to make your workflow smoother. Like you edit the first month to its full length - months = Īnd it automatically corrected the other months in a similar fashion. You can make minor adjustments to the code, and it will follow along quickly. In fact, I copy-pasted the Copilot suggestion above instead of typing it.

With Copilot, I type months = and it auto-fills the entire thing. Which is exhausting to type sometimes, there's no escape, and you have to do it.

I am currently working with time series data a lot, and I have to constantly make lists such as months = I mostly use the Copilot to make long lists and things that are exhausting to type. It makes my life easier but also annoys me at times. There were several occasions when the autofill prompts annoy me, especially when the prompt is not what I wanted. In my typical workflow, I would write the code as if I did not need the assistant, and whenever I feel stuck or lost in ideas, I see the autofill suggestions and fix my code accordingly. We might not use Copilot to write the entire code, but it's good to know the limits of the assistant and all the ways it can help us. I am impressed how it picked up that I would build a KNN model and auto-filled the code accurately. I need to give some context, and the assistant did all the heavy lifting. In the next test, I just wrote a function def knn(): and the Copilot did all the autofill for the next part of the code.

Webstorm copilot