The raw telco data is not the best data set to demonstrate on

[mbkupfer]

Hi, while going through the tutorials, I've noticed on multiple occasions where the functions don't end up having any effect since the data set is already pretty clean. This makes some of the demonstrations a tad bit confusing as they end up being a no-operation procedure. May be beneficial if the raw data set started out more, well raw.

Example 1: In the load data notebook, there are cells that create transformation on null data. None of the data in the file has any nulls though.

Example 2: In the exploratory notebook at the drop_outliers definitions, outliers get dropped in only 1 of all 6 possible drops.

[sfc-gh-cbaechtold]

@sfc-gh-pjain flagging for your awareness

[sfc-gh-pjain]

The purpose of this guide was to make it easy for everyone to get started with snowpark (setup and easy ML engineering) and so made the data set super easy for everyone.

You can use the same notebooks on more robust dataset.