FUTURE WORK

While the project might be a complete working system for now, there are more functionalities that can be added to the system to make it more accurate and reliable.

Firstly, real data can be gathered from different cafes and pubs to create a more reliable training data set to make the predictive models more accurate.
Secondly, it is possible to get customized recommendations based on a user's feedback to the song selection in the playlist. This can be achieved by recording the pressing of the skip button on the Spotify player. The feedback can then be added to the machine learning models which will further streamline and make future predictions more accurate according to the user's preferences.
Thirdly, more selection criteria can be added to the machine learning models such as liveness, loudness, etc. Increasing the number of parameters to the Spotify API might result in a more accurate recommended song list.
Lastly, volume adjustment can be incorporated into a feedback loop to automatically control the volume of playback with respect to the crowd volume in the room