Analysis done on different data sets using Hadoop Tools This word doc contains the screen shots and the commands used to perform a sentiment analyze on the subset of tweets of Mr.Trump The tweet file is a json file which was initially cleaned using Python
The biggest issue of this analysis is that many tweets did not have a sentiment due to lack of rating from the words used in the tweets. Most of the tweets do not have a rating because of the dictionary that we have used which do not contain some words. To get a better understanding of the sentiment of these tweets we should first analyze the words present in tweets and then try to create a dictionary containing all those words with their scores.
Also, sentiment analysis could be done together for a phrase by using a dictionary which a=has scores for each different phrase used. This gives a different and some more clearer understanding of the polarity of the tweets received.