COVID-19 Instagram Analyzer

This project involved designing and implementing a data analysis pipeline to analyze COVID-19-related behaviors and sentiments on Instagram. Our team successfully created a hashtag network that identified frequent co-occurrences of ideas among users. To gain deeper insights, we applied graph clustering techniques and centrality measures to analyze the network. Additionally, we utilized a fine-tuned BERT transformer neural network for user sentiment analysis, leveraging Google Cloud GPUs for efficient processing. Our project secured second place at the STEM Fellowship’s 2020 National Undergraduate Big Data Challenge. You can watch our presentation video and access the code on GitHub. Furthermore, our study has been accepted for publication in the STEM Fellowship journal.