-
The Art of Information Refinery at Cue
Track: Applied Data ScienceLocation:Grand Ballroom - Salon A/BAbstract:
With the advent of cloud computing, humans are creating an immense amount of personal information, yet very little of it is being used to personalize anything we do. Cue is a free service that tries to solve this problem by analyzing all of your cloud content (email, contacts, calendar, etc) to predict what you'll need next in your day. This talk will cover the Redis-powered processing pipeline Cue operates in order to churn raw data into useful information. We'll outline the methodologies we use to extract and understand things like relative times, email signatures, definitive phone numbers ("call me at...") and more. We'll also describe how we identify email patterns (flight confirmations, hotel reservations, shipping notifications), automatically turn them in to structured information and store them using our custom read-heavy Lucene clusters.