Travelling to the Stadium
Rather than looking at just raw data (like the title or artist), DFS creates features by following relationships across related data tables—such as composer history, genre characteristics, or performance metadata—and applying mathematical aggregations (e.g., counts, averages, or standard deviations).
Deep feature synthesis: Towards automating data science endeavors johann_sebastian_bach_toccata_fugue_in_dm_by_si...
: A high-dimensional representation of the performance's "darkness" or "tension," often derived from the frequency of diminished 7th chords or the use of specific minor keys like D minor. Rather than looking at just raw data (like