Excavating Big Data associated to Indian Elections Scenario via Apache Hadoop

Abstract

Data is not a new term in the field of computer science, but Big Data is essentially a new word. When data grows beyond the capacity of currently existing database tools, it begins to be referred as Big Data. Big Data posses a grand challenge for both data analytics and database. It has been only in 2013 to 2015 that we humans have created 90 percent of data existing on the planet earth since existence of humans on this planet. The huge technological up gradation in social network, in retail industry, in health sector, in engineering disciplines, in the field of wireless sensors, in stock market, in public and private sector, all has collectively amassed enormous data. This data is very huge in volume, it gets created at very high speed, it may be structured, unstructured, semi-structured or may be in text, audio or video format and most important that it is not totally precise and can be messy or misleading. The central theme of our research work is concerned with handling huge amount of data that is concerned with different formats of elections that are been contested in India. The framework used in this research work is Apache Hadoop. Apache Hadoop framework makes use of Map-Reduce technology which operates in three steps: mapping, shuffling and reduction. Map-Reduce is the same technique which Facebook use for handling its section of “People you may know”. Research paper also discusses the working of Map-Reduce technology with competent examples. Keywords - Big Data, Big Data analytics, elections, Hadoop framework, Map-Reduce.