Spark Meetup: March 2nd 2016

After a longer gap than usual the first Spark meetup of the year is here! We have two great talks and two awesome lightning talks to make up for the wait. The Spark meetup is now sponsored by Capgemini - who are helping us provide enough spaces for everyone who wants to attend. So the next venue will be even bigger...

Title: Intuition is now a science - How Spark enables an advanced Predictive Analytics platform for Business Analysts

Speaker: Donal Daly (CTO Singularities / CEO Data Wranglers Ltd)

Singularities is developing a SaaS application focused on Business Analysts to really know their customer. In this presentation we discuss why we chose Spark over Flink. How we leverage a Lambda architecture with Spark to make the most efficient (and cost competitive) use of hosted resources. Our evolution off using different Machine Learning algorithms and how and why we use FPgrowth from MLLib today, and our future plans to incorporate additional Machine Learning algorithms. We are currently running on Spark 1.6

Bio: Donal is a long time Database Geek, now Strategist, Technologist and Evangelist for Big Data. CTO @ Singularities. 5+ years of Big Data experience as a Technologist & Evangelist for Big Data (Aster Data & Hadoop at Teradata) and from a consumer perspective (Full Tilt Poker). Now focused on Singularities to bring their product to market. Over 25 years systems software development experience including senior roles at Teradata, Informatica and Oracle (15 years).

Title: Serialization Exceptions and Memory Leaks

Speaker: Sam Savage

Abstract: Many will now be familiar with the dozens of ways to solve the dreaded Serialization Exception. Nevertheless why they happen can still be confusing, subtle and leave us feeling uneasy. Some solutions are ugly hammers and some may even cause your application to slow down or even run out of memory. This talk will discuss the details and present some general design tips to keep your applications fast, hack free and exception free.

Bio: Sam's an early adopter of Spark, using it in production applications for two years. He hopes sharing his experiences will ease further adoption.

Lightning talks by Matt Thomson and Dylan Lentini