Why do you need a train/test split for a z-score? It's just telling you if data is extreme beyond some threshold defined by the distribution from which it originated.

Replace the value 12 in the introduction by 1,000,000 and see if you still agree with your premise :)

I'm interested in the Data Scientist position. How can I apply? The position page just says "For more information, visit us at https://kevalaanalytics.com" but no more information is to be found at the homepage. (: