Inspiration
The inspiration behind our project comes from the lessons learned during the COVID-19 pandemic. We recognized the critical need for timely and relevant information to support effective public health messaging. The project aims to bridge the gap between public health officials and individuals, fostering trust and adherence to guidelines.
What it does
Our project is designed to analyze and understand social media conversations, particularly tweets related to COVID-19. It leverages data collection and a classifier to label users as 'pro-science' or 'anti-science.' This analysis sheds light on the willingness of individuals to accept scientific evidence and trust public health officials, offering a valuable resource for improving public health messaging.
How we built it
We took datasets of examples of health experts (representing pro-science) and pseudo health experts (representing anti-science) and used their tweets as guidelines to label participant user's tweets. Our code is in Python, incorporated into the existing labeler, and compared with the existing classifier. We also used OpenAI API and pandas library.
Challenges we ran into
Our journey was not without its challenges. We faced issues with data collection, model development, and interpretation of results. Ensuring the accuracy and reliability of the classifier was a significant hurdle. Additionally, handling a large dataset requires laptops with large main memories, which we did not have. Since the existing classifier could not be ran locally on our devices or uploaded to cloud, we had to separate our project from parts of the existing model and analyze separately.
Accomplishments that we're proud of
We're proud of successfully creating a classifier that can label users as 'pro-science' or 'anti-science.' This tool has the potential to make a real impact on public health messaging during pandemics. Our team's dedication and collaboration in overcoming challenges have been a significant accomplishment.
What we learned
Throughout this project, we gained valuable insights into the complexities of public health communication during pandemics. We learned about the power of data analytics, the significance of accurate labeling, and the importance of trust in public health officials.
What's next for Health Conversations during the Pandemic
The journey doesn't end here. The next steps involve further refining the classifier, expanding the dataset, and conducting more in-depth analyses of conversation threads. We aim to provide even more accurate insights to public health officials to enhance their messaging strategies during future pandemics. Additionally, exploring the impact of external content, such as URLs, on engagement remains a priority.
Log in or sign up for Devpost to join the conversation.