Exploring ethnic music collections

A system for exploring ethnic music libraries geared towards the general public

The challenge

Ethnic music collections provide vantage points into language, customs and social interactions of different communities. Systems that allow researchers, ethnomusicologists and musicologists these music libraries are numerous, whereas tools that cater to the general public are few and far between.

A solution

Dr. Tiago Tavares and I devised an ethnic music collection exploration system that comprises a minimalist web-radio interface and a recommendation system.

The interface displays track title and duration, and elapsed time, and allows users to play/pause a track, navigate through it via a slider, like a track or skip it. The recommendation algorithm is based on the premise that listeners that like a track would probably enjoy music that is acoustically similar to the said track. If the user likes a track, the next recommendation will be acoustically similar to the previously played track. Conversely, if the user skips a track, the subsequent recommendation will not be similar to the played track.

A/B test

We conducted a blind A/B test to determine if the recommendation algorithm could be employed by the general public to explore ethnical music collections. The Música das Cachoeiras dataset was used in the study.

A website hosted the system for 30 days. The study was advertised using social media to encourage the participation of individuals interested in music.

Each access to the system was labeled as a different session. Within each session the subject was randomly assigned to either the proposed recommendation algorithm or a baseline algorithm that makes random recommendations. All interactions were recorded for further analysis.

Data analysis results suggest that the proposed algorithm could evoke better user experiences than the baseline algorithm. Please refer to Ethnic music exploration guided by personalized recommendations: system design and evaluation spaces with anchors: design and evaluation, published on Springer Nature Applied Sciences for further info.