Sentiment Analysis of Tweets using Apache Flume andSpark

Abstract

Throughsocialmediatheuserscansharetheirthoughtswithfriends,family,andcolleagues,anditalsogivestheuseraplatform
totalkandcommunicateontheirfavoritetopics.This“unstructured”conversationcangive businesses valuable insight
into how consumers perceive their brand, and allow them to actively makebusinessdecisions to maintain their
image. With a rapid increasing of data of sentiments in social media on web has leadtheresearchers into increased
interests regarding opinion mining and sentiment analysis. However, SentimentAnalysisis now considered as a Big
Data task due to the large amount of social media available on theweb.
To find a technique such that it can efficiently perform sentiment analysis on big data sets was the main
focusofthis research. In this paper, Hadoop Apache ecosystem’s data ingestion tool was used to perform
SentimentAnalysison the large sets of data consisting tweets and stream processing with Spark. Using this technique
theexperimentalresult shows very good efficiency in handling big data sets ofsentiment