Original event: http://www.meetup.com/Data-Science-for-Sustainability/events/216522532/
Zico Kolter, Chief Data Scientist, C3 Energy
C3 Energy is a leading provider of cloud-based analytics software for the energy sector. This talk will first provide a short but broad introduction to the work going on at C3, and several of the problems we are investigating. Second, we will also look into substantial depth at a single area of research: producing probabilistic forecasts for energy systems, spanning demand forecasting, price forecasting, and renewable forecasting. In this space, C3 Energy recently participated in the Global Energy Forecasting Competition (GEFCom 2015), where we obtained the highest number of top-ranking entries (in the top 5 submissions) over all teams. We will describe our approach in detail, which involved an alternating direction method of multiplier (ADMM) approach to multiple quantile regression, plus extensive automated feature selection and generation.
Zico Kolter joined C3 Energy as a Chief Data Scientist in 2014. He is a faculty member in the School of Computer Science at Carnegie Mellon University, one of the leading computer science departments in the world, and his research focuses on machine learning and data analysis for energy systems. He has worked on problems ranging from energy analysis of homes and buildings, probabilistic forecasting of energy
demand and generation, wide-area control of grid generation, and assessing asset risk from weather events in distribution networks. He received his Ph.D. in computer science from Stanford University in 2010, and was a postdoctoral researcher at MIT from 2010-2012.
Past talks >