Fast Analytics On Big Data

Explore fast analytics on big data in this 50-minute conference talk from GOTO Aarhus 2014. Dive into an open-source platform for in-memory distributed data processing, capable of handling datasets from 1K to 1TB without code changes. Learn about state-of-the-art predictive modeling and analytics techniques that are significantly faster than disk-bound alternatives and R. Discover how to run R expressions on tera-scale datasets and manipulate data using Scala and Python. Gain insights into the platform's coding style and API that enables seamless scaling from laptop to 100-server clusters. Examine topics such as MapReduce, data layout, chunk alignment, and Spark integration. Witness a live demo and participate in a Q&A session to deepen your understanding of this powerful big data analytics solution.

Intro
Big Data Work
MapReduce
Scholar Version
Map
Count
Set
Group Buy
Uniques
Limitations
Strengths
How it works
Data layout
Conceptual view
Chunk
Alignment
Summary
Demo
Spark Integration
Questions