Use of machine learning is not starts with preparing model for execution. It need first to know the problem. It means ML is tool to solve the problem so for that purpose we need to well define the problem. On that basis we will do data collection, data cleaning, feature extraction, learn to perform task, predict output.
After defining the problem and data manipulation. Data is ready for next process the making the machine to learn the task. There we have to train the x% of the data and y% we will use for the testing to check the prediction accuracy.