It was a dark Saturday afternoon with my fellow data science workshop members when we were trying out different parameters while running a Titanic ML tutorial in a Jupyter notebook. Can we improve our model by using a different model or cleaning missing data in a different way?
In Azure Machine Learning Studio you can create a machine learning model by dragging and dropping modules and setting their parameters. You can use it for free. Azure Machine Learning Studio is a good way to start learning about ml because it has a large number of template models in the gallery that you can study or customize to your problem.
The exam is going to retire in 2019-06-30, I think it involves too many programming languages. Most of us choose between python and r but the exam requires understanding the basics of an R script python and other script languages.
I recommend the exam ref book and getting to know the Azure Machine Learning Studio studio.azureml.net to give structure to studying ML before you start working on your own projects.