Sentiment analysis using tweets

Inspiration

Disasters are serious disruptions to the functioning of a community that exceed its capacity to cope using its own resources. People use twitter for help from government authorities. The sentiment analysis is an important way to identify how people, government authorities respond to a disaster

What it does

The model identifies people's sentiment from their tweets.

How we built it

We used variety of machine learning, deep learning and transformer based methods. We used machine learning algorithms viz

Logistic regression,
Random forest
Decision Tree We also tried Deep Learning methods such as
CNN
Bidirectional GRU
Bidirectional LSTM We used transformer-based approaches like BERT. ## Challenges we ran into We need to combine all datasets and remove the NaN values. We also removed the unimportant columns such as 'num'., 'sentiment', 'timestamp'. We also cleaned the data, removed stopwords, ## Accomplishments that we're proud of We got good accuracy, precision, f1 score. ## What we learned We We learned various NLP algorithms. We also learned how to work in a team and coordinate among ourselves. ## What's next for Sentiment analysis using tweets Scale the model so that it can be implemented on other social media sites such as Facebook, Reddit, Instagram, etc