Data Talks with Dataiku and Pandora: Communicating Data Insights | San Francisco

About this event

Join Dataiku for two discussions on data, covering beer and music!

Talk #1: Data, Beer, and Portland: Recommender Systems 101

Guilherme de Oliveira, Data Scientist at Dataiku, will provide a technical introduction to recommender systems. As a beer aficionado, he built one as he was leaving for a trip to Portland, Oregon, to find out which local beers were most likely to suit his tastes. Guilherme will explain the step-by-step process and discuss the respective accuracy of the results found through the most well-known methods.

Talk #2: Music recommendation algorithms at Pandora

For over a decade, Pandora has been gathering detailed musical data from both Pandora listeners and in-house domain experts. Recommendations based on these data now drive nearly every aspect of the user experience: from every track played in radio, to playlist auto-complete, to station or playlist suggestions in Browse. Claire Dorman, Senior Scientist at Pandora will discuss Pandora’s music recommendation engine, describing some of the many individual algorithms and how and why they are combined into a single recommendation ensemble. She will also present some snapshots of musical trends as seen through Pandora data.