BIG DATA ANALYSIS ON YOUTUBE USING HADOOP and MAPREDUCE

Abstract

We live in a digitalized world today. An enormous amount of data is generated from every digital service we use. This enormous amount of generated data is called Big Data. According to Wikipedia, Big data is a term for data sets that are so large or complex that traditional data processing application software is inadequate to deal with them .Big data challenges include capturing data, data storage, data analysis, search, sharing, transfer, visualization, querying, updating and information privacy. Google‘s video streaming services, YouTube, is one of the best examples of services which produces a huge quantity of data in a very short period. Data mining of such an enormous quantity of data is performed using Hadoop and MapReduce to measure performance. Hadoop is a system which provides a reliable shared storage of such huge datasets on the cloud and also provides an analysis system. The storage is provided by HDFS (Hadoop Distributed File System) and analysis by MapReduce. MapReduce is a programming model and an associated implementation for processing large data sets. This paper presents the algorithmic work on big data problem and its optimal solution using Hadoop cluster and HDFS for YouTube dataset storage and using parallel processing to process large data sets using Map Reduce programming framework. In this paper, we solve two problem statements using the YouTube dataset – top 5 video categories (genres) with the maximum number of videos uploaded and top 5 video uploaders on YouTube. A particularly distinguishing feature of this paper is its focus on analytics performed in unstructured data, which constitute 95% of big data.