Of all areas of Trend Forecasting, I find music might be one of the hardest to work on. There are so many different groups, genres, artists, styles... Or maybe, and precisely because of this, it should be easier to forecast the next "sound", as the more information, the better you can extract conclusions.
