I have a dataset if payment card transactions $($ purchase and ATM cashwithdrawal$)$ that considered legit $($ approved $),$ I want to use unsupervised machine learning to get insight from this dataset sample to analize data and detect suspecious behavior or possible fraud $,$ how to do it step by step using Python $?$ explain the methods used in details please $.$