Exploring Big Data Analytics in Social Networking Sites

Over the last few decades, humongous amount of data is being created by humanity owing to heavy usage of social networking sites. Explosive growth of social networking sites such as Facebook, twitter, YouTube WhatsApp generates massive amount of unstructured data. Major challenge faced in this research domain is to convert this huge community data into data which can be analyzed and secured and easily available. In this paper, various tools and frameworks like Hadoop, Spark, Map Reduce, HIVE and MongoDB have been studied in detail for their suitability to analyze large volume of variety of data. The tools have been studied in view of the three characteristics of data. First characteristic is based on nature of data i.e. whether data is historical or live streaming. Second characteristic considered is mode of analytical application i.e. batch or online. Third characteristics is data type whether data is structured, unstructured or semi-structured. From review of the papers, it is observed that in case of historical data, structured type and batch mode analytical application, HIVE is most appropriate on Hadoop framework. In case of live streaming data and online analytical application, MongoDB on Spark framework is appropriate. In case of semi structured and historical data, Mapreduce is appropriate in Hadoop framework. For analyzing data generated from social networking websites, Hadoop with HIVE is observed to be most suitable. This paper discusses commonly used tools and their comparison on these three characteristics.