Turning Up The Music With iHeartMedia
With over 850 live broadcast stations in 160 markets, iHeartMedia partnered with Aviana to optimize music programming operations using AI-based technologies. The results include an Aviana model that predicts song popularity, building recommended playlists tailored to specific genres, markets, and song mixes, increased efficiencies in the time it takes to build playlists, and reduced labor costs through the improved management of music programming for multiple stations.
Listening To The Data
Model – Predict Today’s Song Popularity:
Utilize song and artist-level local & national data from a variety of in-house and 3rd-party data sources to create a song popularity estimate.
Use Cases & Benefits:
- Provide more stable music data that is less subject to “sample bias”
- Provide song poplularity metrics for more songs each week
- Provide song popularity metrics for all stations and markets
Creating The Perfect Playlist
Model – Predict Future Song Popularity
Utilize song and artist-level local & national data from a variety of in-house and 3rd-party data sources to predict song popularity in the future, to derive recommended play lists tailored to each station.
Use Cases & Benefits:
- Predict the song “lifecycle” curve to supplement gut-feel with data.
- Identify what are the best songs to add to rotations.
- Identify which songs to move into a power position.
- Get the most out of a popular song and not move a song out of rotation too soon. Identify which songs should be moved out of rotation.
Predicting The Next Big Hit
Model – Early Indicator
Predict early in the song lifecycle that a song will be a hit or not.
- SPSS Modeler was used in all exploratory data analysis, data wrangling and the creation of models in the projects.