/Tweepy-Sentiment-Analysis

Scraping twitter using Tweepy and exploring different preprocessing techniques and models for sentiment analysis.

Primary LanguageJupyter Notebook

Tweepy-Sentiment-Analysis

Twitter is an enormous source of information on variety of topics, the extracted data can be used for useful insights, information extraction, sentiment analysis and much more! In this project, I have extracted tweets with the help of Tweepy API. Next I have explored a few preproccessing techniques and models for sentiment analysis. This project is still under progress, the primary objective of this project is information extraction, which is yet to be implemented.

Libraries used:

  • Tweepy
  • NLTK
  • TextBlob
  • Spacy
  • Sci-kit Learn
  • Pandas
  • Numpy
  • Tensorflow
  • Keras
  • XGBoost

    • Files:

    1. appCredentials.py : to specify access tokens for twitter api
    2. Stream.py : Classes and methods to stream tweets and attributes from twitter api.
    3. preprocessor.py : Preprocessing techniques:
      • Stemming
      • Lemmatizing
      • Subjectivity and Polarity
      • Frequency Distribution of words
      • Count Vectorizer
      • TF-IDF matrix
      • Text to Sequence
      • POS tagger
      • Name Entity Recognizers
    4. models.py : Models explored:
      • SVM classifier
      • Naive Bayes Model
      • XGBoost classifier
      • LSTM
    5. sentiment_analysis.ipynb : Exploring the various techniques listed in the modules above