This repository pulls 50,000+ Tweets each week that mention US States. With this data, each state is ranked in terms of sentiment and a state word cloud is created.
The sentiment models are located here: https://github.com/ghadlich/sentimentmodels
| Model | Notes |
|---|---|
| Composite | Summation of all models |
| Distilbert | Trained on Kaggle dataset of ~1,600,000 tweets |
| Flair | Pretrained Model |
| Naive Bayes | Trained on NLTK Tweet Dataset |
| TextBlob | Pretrained Model |
| Transformer | Pretrained BERT model from Hugging Face |
| Vader | Pretrained Vader model from NLTK |
