Featured image of post Music development with data

Music development with data

Music collects data and we can share that to develop better products

Marc has always been interested in learning the piano and recently he downloaded a new app called “Happy Keys” which boasts a collaborative and gamified approach towards learning the piano. He realises pretty soon that the rave reviews of the app were well-justified, not only does the app teach the piano but it also encourages him to improvise and create new tunes that sound good to him. Three months into using the app he gets a notification that showing him the work of Juliette a budding Jazz pianist who has been sharing original compositions for multiple years. Marc listens to the song suggested by Happy Keys and loves the intricacy of the music and contacts Juliette to see if he could have a few 1-on-1 mentoring sessions with her to learn from her technique. Juliette is excited to meet up but she has a better idea, she would like to also gather some other friends to jam together and create something together, she explains that the level of playing does not matter but it is about making music together by inspiring each other. Marc agrees and a week later is glad he did, they gather using a “Happy Keys” Jam Now feature to jam together online.

A week after this session Marc and Juliette both receive a notification from “Happy Keys” telling them that a fellow user has identified that their music has a catchy quality to it based on analysis with multiple song samples and there could be potential there for improvement. They could choose to listen to the sample of their music which caught the interest of the algorithm and an analysis on genres and similar tracks for inspiration. They will not need to pay anything for this insight, however if they wish to continue using the algorithm to understand these analytics and build their own tune with the help of the algorithm then they will have to commit 0.2% of any profits to the user who build the original algorithm. (BUG: Simplify payment scheme)

Marc and Juliette accept after all 0.2% of whatever they may gain is a small price to pay but they are also amazed it only costs this much. They review the analysis and use the algorithm to build a very catchy jingle and place it on the Happy Keys marketplace and place it available for short video creators. Within a month they notice that their jingle was downloaded 10 times netting them a tidy sum of €20 each. (BUG: ??? ECONOMICS?)

A year later Marc looks back and finds himself very pleased with the outcome of a small hobby, he has enjoyed collaborating with many people and created some inspiring music but many people have also reached out to him to be mentored and some even view him as an expert in making jingles. Not counting the money he has made from his new course on making jingles he has made about €300 on his contribution to the music he made with Juliette and her friends.

Pulling back the curtain

So I guess it is becoming clear that Happy Keys was not just any other Piano learning app. Happy Keys was collecting Marc’s piano strokes and quantifying it for use in other algorithms, other users are able to connect to this dataset however they do not see which User account is responsible for a specific piece instead algorithms can be built on this dataset to differentiate pieces. An unknown user actually built an algorithm that first distilled Marc’s improvisation style and matched Marc with Juliette to set up that initial connection. And of course when they were jamming together there were other algorithms that could be run against the music, one of these algorithms had quantified what a catchy tune was with about 80% success and suggested this back to Marc and Juliette who proceeded to gain success with their catchy tune.

However still the question remains of how could the creator of the algorithm that identified catchy tunes afford to only request 0.2% of the profits (about €2.4)? Well the creator of the algorithm spend quite some free time building out the algorithm however they did not have to market the algorithm, instead all they needed to do was make it available on the Happy Keys platform and the Happy Keys platform automatically ran the algorithm based on some pre-selection algorithms to see if the music was catchy. Happy Keys only charged the creator of the algorithm about 5% of their profits to constantly run the algorithm. However this algorithm made about 400 suggestions in the past year and with an 80% success rate the creator of the algorithm netted about €760 after paying Happy Keys hosting charge for a hobby algorithm.

Photo by blocks on Unsplash

Licensed under CC BY-NC-SA 4.0
Last updated on Nov 25, 2022 08:30 UTC
comments powered by Disqus
Built with Hugo
Theme Stack designed by Jimmy