MediaMath Masters Massive Data with Fast, Scalable Reporting and Non-traditional Analysis
August 5, 2014 – MediaMath, creator of the TerminalOne Marketing Operating System™, is committed to empowering marketers the world over with its technology, which is built by a team of world-class engineers with unique perspectives on big data.
In a series of recent blog posts on MediaMath’s new Developer Blog, MediaMath engineers weigh in on how they are processing, analyzing, and optimizing MediaMath’s massive amounts of data, making it easier for marketers to gain more actionable insights.
In his blog post “Building faster, scalable reporting with Hadoop-Impala,” Ram Narayanan, Senior Director of Database Architecture and Operations, explains how MediaMath is able to provide half-hourly reporting to clients of their ad attribution data, a problem that requires joining 30 days of ad impressions (40 billion records) with one hour’s worth of click-event data (20 million records).
Ian Hummel, Director of Data Platform, investigates how best to optimize Hadoop’s test and development environments to match the production environments of your AWS Elastic MapReduce (EMR) cluster.