First one must collect the tweets from twitter and the number of tweets to be collected range from 50,000 to 100,000. After collection of tweets they need to clean the data i.e, remove extra spaces, blanks, punctuation i.e, Clean and tokenize the data. Then calculate the frequency of the most common terms. Look for insights about the search term(search terms will be "Hilary Clinton", "Donald Trump") given the most frequent terms. later their tweets are to be compared whether they are positive tweets or negative tweets and with the help of that data they must plot a graph representing whose tweets are more .