Saturday, August 27, 2011

How to Double Your Klout Score in Less than Two Weeks

The Klout Score is the current standard measure of online influence. Klout takes in terabytes of data of social interactions from across the web, from Facebook to Twitter to Youtube. Using large scale analysis and distributed computing, Klout is able to perform predictive analysis to understand when people create content, and how others react to that content. The Klout approach aggregates signals, content, and social graphs from various platforms into a single Hadoop cluster. Once the data is aggregated, Pig and MapReduce jobs run machine learning algorithms that produce metadata for users and their friends. This metadata is loaded into Elastic Search as well as HBase for consumption. Although this provides a robust and scalable data processing pipeline, a couple key areas that contribute to success are 1) understanding social media knowing what questions to ask of the data and 2) building feedback loops into your content so Klout learning models improve over time.


So, there you have it.  Blog more.

No comments:

Post a Comment