The University of Auckland

Project #18: FluxMusic: Crowd-sourced Music Suggestions using Machine Learning

Back

Description:

Many public venues of which music is an important addition often hire DJs and the like to provide a selection of music and tie it together. More often than not, patrons of the venue will be constantly suggesting songs to be played to the host and often results in juggling devices or rejections. An opportunity exists to provide software that would allow venue patrons to choose the music themselves, and for the system to handle the rest.

A web service that hosts playlists which would be editable and accessible by anyone with the password, or perhaps on the same Wi-Fi network. The primary use case would be that a single person can plug their device into some audio system, and anyone else present could go onto this web service via a webpage or mobile app, and queue songs to be played. There are many features that could be included, and our scope is yet to be defined. However, one of the major features we would definitely research would be the implementation of a beat matching algorithm present on the web service so that the playlist flows better, therefore reducing the number of abrupt genre, tempo changes. Another possible major feature we may consider is the ability to stream music from one mobile device to the host device, in case the host device does not have the wanted music selection. The potential incorporation of a natural language processing element for commands and reporting could be another avenue of research. Finally, the Spotify API could also be considered to further expand music availability.

Type:

Undergraduate

Outcome:

Prerequisites

Pre-allocated

Specialisations

Categories

Supervisor

Co-supervisor

Team

Lab

Lab allocations have not been finalised