This project takes you through the following;
- Text analytics
- Text preprocessing
- Topic modelling with LDA
- Dynamic Topic Modelling with LDA
- Topic modelling and Dynamic Topic Modelling with BERTopic
Having conducted this experimentation, I believe the LDA approach is more suited for this project with the use of bigrams to capture the required topics for this corpus.
Feel free to replicate this project
Needed packages can be found in the yaml file
Sentence transformer --> all-MiniLM-L6-v2m
Data could be downloaded from the kaggle-url