r-music is a mobile app that can quickly recommend suitable songs based on the conditions and situations that the user’s need.
With r-music, users can select and enter their condition. Our system classifies each song through song lyrics and song appreciation, combined with machine learning, and uses artificial intelligence to train a model to calculate the similarity between words to provide more accurate songs.
Our initial idea was to not only recommend songs that meet the user's conditions, but also add a social function to allow users to share their moods and music with each other.
We have a total of six members, two are responsible for the database, two are responsible for the back-end algorithm, and two are responsible for the front-end.
Another classmate and I are responsible for the front-end part, and we divide the work according to different interfaces. I am responsible for classifying, inputting conditions, and displaying the music list, the other teammate is responsible for music playback and my favorite list. Here we used Android Studio for our app development.
Users will first select the song category they want to find on the homepage, and then select and enter the conditions for the song they want to find on the next page.
Based on the criteria entered by the user, the app will suggest the ten most relevant songs, and the user can also refresh the list to find other related songs.
Once users select a song, they can listen to it and add it to their favorites so they don't have to search for the same song again in the future.
We found eleven users to help test our first version product. The majority users said their overall experience of the app is nice. In addition, 90% of the users indicated they were satisfied with song recommendations.
During the interview process, users gave us some valuable feedback that helped us improve our products.
Based on our user feedback, we found that our products lacked descriptions and instructions, causing confusion for some users.
Therefore, by adding text descriptions we hope to better guide users on how to use the product.
Take the action of deleting a song from a playlist as an example. In our initial design, we only designed the delete option to be presented in a sliding manner, but this might cause some people who are not familiar with technology products not knowing how to operate.
Hence, we should provide users with more than one option on how to perform an action to enhance the user experience.