Preparing the dataset is 99.9% of pretty much all ML and data science.

Just like all software engineering is maintenance, all DS is data cleaning.