Twitter_Sentiment_Analysis

Sentiment analysis on three different datasets using Lexicon and Rule-based sentiment analysis tools

Data Collection:

1. #Strangers Things
2. #Weather
3. #USAirlines

Pre-Processing:

1. Conversion of Tweet texts into Lower-case
2. Tokenizing the sentences (using NLTK Tokenizer)
3. Removing Twitter Usernames (using Regular Expressions)
4. Removing Tweets which contains URLs (using Regular Expressions)
5. Using Stop Words available in English language dictionary
6. Joining Meaningful words after splitting them

Tweets Labelling (Data Coding):

• Done using 'CrowdFlower'

Sentiment Analysis:

1. SentiWordNet (Lexical Resource used for 'Opinion Mining')
2. VADER - Valence Aware Dictionary and sEntiment Reasoner (Lexial and Rule-based Sentiment Analysis tool - Commonly used to analyze sentiments expressed in Social Media platforms)

Performance Evaluation was done, inferences were made and results were discussed.