Feature Engineering is a powerful technique to augment your data set with 'derived facts' based on the current data sets. These features typically fall into two categories:
- Structured Data (Numerical fields, dates, master data)
- Unstructured Data (Textual Data sets)
The fundamentals are the same - you are adding new columns of data from which the models can use to help see if there is correlations/patterns that help one answer your question at hand.
Structured Data (Numerical fields, dates, master data)
Unstructured Data (Textual Data sets) are typically handled with more advanced techniques like natural language processing (NLP)