Performing NLP for political tweets

• Course: Machine Learning Applications
• Final grade: 9.8/10

Tools and covered topics

• Google Colab
• Rule-based and embedded transformers
• Topic modelling: LDA

Expected Results

The final result should represent the language used by each political party through visualizations, leading to a clear view on Spanish Political picture.

Table of contents

1. Introduction
2. Data Extraction: Twitter scraping and loading
3. Training set: Preprocessing and Vectorization of tweets
4. Testing set: Preprocessing and Vectorization of tweets
5. Training and evaluation of the ML Model
6. Comparison against pre-trained transformers (positivity and hate speech in rule-based and embedded transformers)
7. LDA: topic modelling with gensim
8. More data visualizations on political tweets
9. Final conclusions

Contributing

Contributions are welcome! Please fork the repository and submit a pull request with your changes. Feel free to customize it further to match your repository's specific details and needs!