Sentiment Analysis of Real Time Twitter Data Using Big Data Approach

Online social media is an important platform over the globe to share once opinion. People go for online social media as it is easy to convey their opinions, to have an up-to-date knowledge about the ongoing trends on a daily basis. Twitter is one of the biggest and renowned online social media gets a large number of tweets each day on various topics. This huge volume of raw information can be used for Social, Industrial, Economic, or Government approaches by arranging as per our need and processing. Hadoop is the best tool for analyzing the twitter data as it processes the huge sets of data in parallel. Since twitter contains a variety of opinions on various topics it is necessary to analyze these opinions to know the customer behavior. Analysis is done using Hadoop and its ecosystem to know the polarity of every tweet. In this paper, we have discussed how effectively sentiment analysis is performed on the tweets gathered from twitter using Big data approach. We have used Apache Flume to stream ongoing twitter data into the Hadoop Distributed File System (HDFS). The pig scripts are written to extract tweet from the raw nested twitter data. The dictionary based method is used for sentiment analysis and tweets are classified into three groups such as positive, negative and neutral.