Thursday 20th September, 2012
1:40pm to 2:20pm
Musicmetric collects and analyses online data to present a complete picture of how people are engaging with artists online. Data is collected and combined from social media (activity timeseries, demographics), peer to peer filesharing (geographic fan activity) and reviews (NER, sentiment analysis). When we started processing increased volumes of this public data, along with private data sources such as iTunes and Spotify transactions using Hadoop we faced a new challenge; how to serve the data for hundreds of thousands of artists with our RESTful API. Of the myriad of key value stores we chose Project Voldemort. In this talk we describe our data processing pipeline and how we came to choose Voldemort.
Sign in to add slides, notes or videos to this session