David Smith, Revolution Analytics | BigDataSV 2014
summary
Company announces Revolution Analytics on AWS Marketplace and explains how R scales for production by adding an xdf on-disk data type that keeps datasets on disk and streams rows so algorithms perform incremental updates. Engineers implement streaming updates for logistic regression and trees, delivering millisecond turnaround and Hadoop/YARN and Spark integration so analytics run where the data lives. The claim: large datasets no longer must fit memory and production workloads can run in cloud or Hadoop without extracting terabytes.
AI-written summary · about
chapters (12)
- 0:00Welcome to BigDataSV 2014
- 1:05Revolution Analytics overview
- 3:31Enterprise R for data science
- 4:35Scaling R for big data
- 6:20Analytics in Hadoop environments
- 7:25Revolution on AWS announcement
- 9:10Cloud barriers and opportunities
- 10:12Visualizing and democratizing data
- 12:18Market pressure and growth
- 13:42R integration with Spark and YARN
- 15:06Cloud and devops impact
- 16:32Industry maturity and big data value
