The Fifth Elephant 2013

An Event on Big Data and Cloud Computing

Extracting consumer trends in real time using 100 billion tweets.

Submitted by Pankaj Risbood (@risbood) on Wednesday, 27 March 2013

videocam_off

Technical level

Intermediate

Section

Analytics and Visualization

Status

Confirmed

Vote on this proposal

Login to vote

Total votes:  +59

Objective

Connected consumers express everything they feel about products, services, brands in social media. How consumer feel and engage with products is very important for retailers to make better merchandizing decisions.

In this talk we will outline one such system we built to analyze 13 months of twitter data amounting to ~100B tweets.

Description

We will outline a system to extract volume, sentiment, geo trends and related words about any arbitrary topic (defined as a boolean query) from a corpus on 100B tweets increasing at ~500M a day.

Speaker bio

Pankaj Risbood is Director of Engineering at @WalmartLabs where he leads social media analytics effort. Prior to @WalmartLabs Pankaj spent 6 years at Google leading various efforts in cloud computing, enterprise and speech processing. He was Member of Tech Staff at Bell Labs where he specialized in optical and IP networking. Pankaj has co-authored 16 issued patents and several research papers in premier conferences. An alumnus of IISc, Pankaj is a avid hiker and a marathon runner.

Comments

  • 1
    t3rmin4t0r (@t3rmin4t0r) 5 years ago

    Is the platform in any way open source or based on a white-paper available from Walmart?

  • 1
    Pankaj Risbood (@risbood) Proposer 5 years ago

    The solution developed uses open source technologies -- hadoop, hive, hamake, solr and mupd8.

Login with Twitter or Google to leave a comment