In machine learning workflows, irrelevant features can weaken models. Using df.drop ensures only relevant columns remain, leading to cleaner training data and better results. https://docs.vultr.com/python/third-party/pandas/drop