View Proposal


Proposer
Habte Lejebo Leka
Title
Social Media Sentiment Analysis and Visualization System
Goal
Description
The project aims to develop a system for analysing public sentiment in social media posts toward a selected topic, product, or brand. Students will collect and preprocess social media text, handle characteristics such as informal language, emojis, and abbreviations, and develop suitable approaches for sentiment classification. The system will classify posts into positive, negative, and neutral sentiments and provide an interactive dashboard for visualizing sentiment trends and analysing public opinion.
Resources
Atandoh, Peter, Zhang Fengli, Daniel Adu-Gyamfi, Habte Lejebo Leka, and Paul Hakeem Atandoh. "A GloVe CNN-BiLSTM Sentiment Classification." 2021 18th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP). IEEE, 2021: 245-249.Dataset: Twitter Sentiment Analysis Dataset Kaggle Link: https://www.kaggle.com/datasets/kazanova/sentiment140
Background
Url
Difficulty Level
Moderate
Ethical Approval
None
Number Of Students
2
Supervisor
Habte Lejebo Leka
Keywords
sentiment analysis, social media analytics, natural language processing, text classification, machine learning, deep learning
Degrees
Bachelor of Science in Computational Sciences and Software Engineering