Adventures (And Misadventures) in Automated Insight Discovery

Michael Kim | Outlier AI

ABOUT THE TALK

Today, companies are generating increasingly large amounts of data about more and more aspects of their business. And even with the incredible advances in data solutions (e.g., data collection, storage, query, visualization, prediction technologies), smaller and smaller fractions of this increasing volume of data is being analyzed.
 
One of the manifestations of this growing gap in analytics is the anti-pattern of running "fire drills" when urgent exceptions are found in the data (e.g., an unexplained drop in revenue, or spike in costs, etc.). Often, these urgent exceptions are the result of relatively small issues that have gone undetected for some time. With automated insight discovery, we have the potential of finding these problems while they are still small and addressing them before they have a larger impact.

While the promise of automated insight discovery is alluring, this is a very challenging problem. We will describe some of these challenges and also our particular approach to insight discovery.

Download Slides

michael kim

CTO | Outlier AI

Michael is the CTO of Outlier AI. Previously he was the Director of Engineering & Head of Research at Alt School. He also worked at Google as Software Engineer. He did his Ph.D in Bioinformatics from University of California (UCSF).
Mike Kim

powered by
hakka-labs-logo.png

PO Box 2087
Wilson, WY 83014