Picture credit: Microsoft Bing Co-Pilot Image Generator
One of the problems associated with learning to code is that the initial hurdle of just getting your first line of code to run is pretty high. You need to get comfortable with a command-line driven interface. You need to install multiple packages depending on your particular needs, and the errors messages are hard to deciper.
Google Colab dramatically reduces the initial hurdle. First, Python is pre-installed so you do not need to worry about that. Second, you can annotate like you do in a regular Google Doc (in fact, it was built as part of Google Doc). Third, you can execute line by line and see what happens, which makes debugging a bit easier, or you can run the whole code file. Fourth, it has access to GPU and TPU power for advanced AI modeling.
I wish I had known Google Colab earlier as it really makes coding in Python so much more accessible.
Here is an excellent video walking through running regression in the context of AI modeling. I like it because it covers a series of good data analytics practices such as cleaning data, recoding variable, checking model overfitting and predictive power.