This project was done for our NUS module: BT4222 Mining Web Data for Business Insights.
This project aims to create models to predict sentiments and extract key information and topics from course reviews to help students and professors analyze them with ease. We utilized an open-source dataset from Hugging Face (https://huggingface.co/datasets/kkotkar1/course-reviews/viewer/default/train) and web-scraped reviews on Coursera. Both datasets were then merged and cleaned prior to model building.
The code is separated into 2 notebooks. One contains the data pre-processing and model building for sentiment prediction while the other contains the models for topic extraction.
Done by:
Gui Jun Yao
Audrey Goo
Chan Choon Leng
Carissa Ying
Kuek Yan Ling
Chanell Ng