Schedule: Hadoop & Big Data: Applied sessions

In this session, Hortonworks CEO Eric Baldeschwieler will look at the current state of Apache Hadoop, how the ecosystem is evolving by working together to close the existing technological and knowledge gaps, and present a roadmap for the future of the project. Read more.

This session will draw on numerous customer examples to reveal powerful tips, tricks, and in-depth use cases to show how Hadoop can easily integrate, scale, and analyze important data. Read more.

In this talk, we'll build a complete, scalable collaborative filtering ("people who X also Y") system that is almost identical to what prominent Internet properties use today. We'll talk about model improvements, performance enhancements, and practical considerations. This is a practical talk accessible to all. Read more.

As more companies adopt Hadoop to perform data intensive tasks for large data sets, there is a burning need to make Hadoop available to a broader set of developers. This talk covers two approaches Microsoft is exploring for this purpose: 1. JavaScript interfaces to run Hadoop jobs and 2. web interfaces for Hadoop that let developers write and run MapReduce jobs from any platform. Read more.

This talk will go in details, architecture and challenges of building a recommendation system on a massive social graph. The talk will describe how we applied learning on large datasets using Apache Hadoop and how we scaled millions of reads and writes. We will also showcase Eventbrite's data platform architecture. Read more.

This session discusses financial services use cases and challenges in using Hadoop analytics including long-term storage and analytics of transactions, identifying cross and up sell opportunities by analyzing web log files and customer profiles, value-at-risk analytics, and understanding the SLA issues and identifying problems in a thousands-of-nodes, big-services oriented architecture. Read more.

Hundreds of hours of video recordings culled from multiple cameras. Most of these recordings hold little value as the scene does not change for extended periods of time. For organizations that must keep the original in tact, analyzing these recordings can be very difficult. Using Map/Reduce we can harness parallel processing to identify and tag useful periods of time for faster analysis. Read more.

Map/Reduce has created tremendous interest in parallel programming and big data analytics, but it isn't always the right tool for the job. Many new projects have emerged in this space over the last year including two cluster schedulers (YARN and Mesos) and numerous parallel computing environments. We'll provide an introduction to these new technologies, including some you might not have heard of. Read more.

NetApp collects 250 TB per year of unstructured data from devices that phone home. They need to be able to do ad hoc analysis and build predictive models for device support and cross-sales. We discuss our experiences building a Big Data system with NetApp using Hadoop and HBase to improve customer service, drive sales and develop better products. Read more.