COVIDFact

Analytics of social media discourse related to COVID-19 including trends, topics and trust

Project Organization

├── data               <- Data files for use in training
│
├── models             <- Trained and serialized models (pickle files)
|
├── README.md          <- The top-level README for developers using this project
│
├── covid_classify.py  <- Code for training classifier of COVID/non-COVID posts
│
├── covid_predict.py   <- Code for predicting posts as COVID/non-COVID
│
├── extract_phrases.py <- Code for extracting symptom & treatment phrases
│
├── med_non_med.py     <- Code for training classifier of medical/non-medical posts
|
├── reliable_unreliable_model_trainer.py <- Code for training reliable/unreliable model
│
├── requirements.txt   <- For reproducing the analysis environment (run pip install -r requirements.txt)
│
└── word2vec_train.py  <- Code for generating COVIDFact embeddings from abstracts

Dependencies

• Part of the code depends on the Word2vec Twitter Tokens file (Google Drive link provided due to large size of 4.6GB).
• To install all Python libraries and modules needed, please run pip install -r requirements.txt

---

Repository structure partially based on the cookiecutter data science project template.