Twitter sentiment analyzer
Technologies:
NLTK, Elastisearch and Kibana.
Description:
- A Naive Bayes Classifier is trained on a set of 21k tweets. It is then used to classify live Twitter streams as positive, negative and netural.
- Tweepy (Twitter API) was used to stream live tweets for a topic (input to program).
- Sentiments which were output of the classifier were indexed using Elasticsearch.
- Kibana was used for plotting graphs of sentiments predicted by live tweets.
Steps:
- Clone the repository.
- Run the files in following order
- train_classifier.py
- load_classifier.py
- twitter_stream.py
- train_classifier.py