Follow Datanami:
September 21, 2012

Core Analytics Puts Hadoop to New Use

Datanami Staff

With the rise of Facebook came the rise of social gaming. There was a point at which friends could be found postponing important social functions to harvest crops, water strawberry fields, and collect coins on Farmville. While that particular game has become out of fashion, social media games themselves live on and are changing the face of the video game industry.

Core Analytics co-founder Kimberly Chulis discussed the changing landscape of the gaming industry and its relation to data analytics, specifically with regard to NoSQL databases, in a presentation at NoSQL Now! in San Jose.

The console-based gaming industry follows and has followed the same model for quite some time. Companies make consoles and games and sell them. Consumers play them and either love them, hate them, or feel okay about them. Gamers are only too happy to give their feedback so the developers know what to improve. While the World of Warcraft-based gaming industry follows a slightly different business model, which relies on subscriptions, their developers also get constant feedback from their devoted gamers.

On the other hand, social games are more complex from a data collection standpoint. Since they make money primarily through in-game advertisements and monetized reward systems, they need significantly more insight. According to Chulis, a MySQL database is simply not enough to handle those demands. “The traditional relational databases,” Chulis said “that game developers relied on typically to analyze gameplay, subscriptions, customer feedback, can’t scale to the new demands. That’s why the NoSQL databases are necessary to efficiently store all of this data into rich, actionable player insights.”

To remedy this, Core Analytics, like many other new converts, relies on Hadoop via Cloudera to build a database that can not only handle the, according to Chulis, hundreds of millions of virtual transactions that take place but also the vast amount of social media data that accompanies it.

Catering to a particular gamer, through a process Chulis calls “Actionable Player Insights,” is the goal. While video game companies can get away with letting some people hate their game if enough people buy it, social media games demand a shorter attention span and therefore a less loyal following. In order to make a profit, social game developers need to direct gamers to exactly the right advertisements.

Core does this specifically through JACT Media, a company which specializes in monetized reward systems and in-game advertisements. According to Chulis, their MySQL database could not scale to the millions of coin transactions that took place, nor could it handle any of the communicative data which gave insights into which advertisements customers were likely to enjoy.

Interestingly, Chulis noted that Core was bringing Hadoop capabilities and data into the MySQL relational database.

The gaming industry is careening more toward social media flash games, and Core Analytics, powered by Hadoop, is here to make sure it maintains profitability.

Related Articles

The Algorithmic Magic of Trendspotting

Study Stacks MySQL, MapReduce and Hive

Six Super-Scale Hadoop Deployments

Datanami