This repository contains our solution on analysing the Voice of Customer (VOC) for the module DSA4263: Sense-making Case Analysis: Business and Commerce in NUS.
Given a dataset containing the customer's feedback, we were tasked with finding:
- Sentiment analysis (Is the feedback positive or negative)
- Topic modelling (What topic is the feedback on) Afterwards, using our trained machine learning model, we visualise the change in trends sentiment for the topics.
The repository contains 2 main subfolders, sentiment analysis and topic modelling. And within both folders, there is a BERT and non-BERT related model solution. The main page contains files related to our preprocessing, docker and flask solutions.
Use the package manager pip to install the requirements
pip install -r requirments.txt
With the root directory as the working directory:
docker build -t <image_name> .
docker run -d -p 5000:5000 <image_name>
The website can be accessed by localhost:5000.
Chan Zhen Hao, Benny
Choong Meng Zhun
Goh Jia Jun
Lim Zi Hong
Liu Ting Yen