![]() We glean a lot from your activity with us via thumbs. The challenge is determining what the next best track should be. Many listeners seem to prefer staying on one station most of the time. Listener preferences for their radio stations are complicated. At the scale that we operate, this makes Pandora uniquely equipped to truly understand what a listener will want to hear. We have the largest catalog of annotated songs in the world. This is a highly scalable approach and can be used on any number of tracks that we ingest. A sophisticated machine listening system utilizes computer-based audio analysis and machine learning to help us determine what the next track should be in your listening experience. This remained the name of the project until late 2015 when it was released under its final moniker, Thumbprint Radio.Īt Pandora, we have about 450 characteristic musicological traits for each track. We chose “Project Orange” because orange is the color at Pandora. The name for the product remained elusive. ![]() ![]() With three teams from different areas of the company aligned, the project started off as an investigation with a small group in January 2015. The engineers in the Computational Programming team also thought it was a great idea. All of that didn’t matter to the Product Management team. However, it was raw and it definitely wasn’t scaled. The Data Science team thought that it was an inspired idea. The demo was viewed by a few hundred people inside of Pandora, most of them from the Product and Technology team. It put all their tracks into one highly personalized station. It tangentially addressed a request that listeners had to replay a track that they thumbed up. They created this promotional video to introduce their hack presentation. For most of 3 days, this team hacked their way into creating a station that only played music that a listener thumbed up (i.e. The idea, named “My Thumb Mix,” purported to “create a station that plays only songs that you’ve thumbed up.” The team was composed of 10 people: an analyst, a product manager, a security engineering manager, a data scientist, and 6 engineers from various teams. During our bi-annual hackathon in November of 2014, we were treated to a great idea.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |