VaxLens: Data-Driven Vaccine Insights

Inspiration

The rapid spread of vaccine misinformation inspired us to create a tool that leverages data to fight false narratives. We wanted to build something that empowers public health efforts and restores trust in vaccines by addressing misinformation with timely, fact-based responses.

What it does

VaxLens tracks vaccine-related tweets or bluesky posts, identifies negative sentiment, and extracts trending keywords. It highlights impactful tweets based on likes and reposts and suggests scientifically accurate responses using AI agents. The platform includes a dashboard for visualising insights, creating counter-narratives, and fostering informed public health communication.

How we built it

• Agents: to automate keyword detection, response/post drafting, perform web searches and content prioritisation based on tweet engagement metrics like likes and retweets.
• Data Collection: Leveraged APIs to fetch vaccine-related tweets and applied natural language processing (NLP) models to classify sentiment.
• Keyword and Trend Analysis: Used algorithms to detect frequently occurring terms in negative tweets and identify high-impact misinformation.
• AI-Powered Responses: Implemented agents to generate informative, empathetic counter-tweets tailored to address misinformation.
• Interactive Dashboard: Built with Streamlit, displaying flagged tweets, suggested responses, visual analytics, and data-driven insights.
• Visualisation: Incorporated custom graphics and charts to communicate insights effectively.

Challenges we ran into

Noise in data: extracting meaningful insights from noisy, unstructured social media data was a significant challenge. Accuracy of sentiment analysis: ensuring tweets were correctly classified as negative, neutral, or positive required fine-tuning NLP models. Balancing tone: crafting responses that were not only accurate but also empathetic and approachable was critical to engagement. Scalability of the project: processing and analysing large volumes of tweets in real-time required optimising workflows and algorithms.

Accomplishments that we're proud of

• Successfully creating a system that combines sentiment analysis, trend detection, and AI-generated responses to counter misinformation.
• Building an intuitive and visually appealing dashboard that empowers users to address vaccine-related concerns effectively.
• Designing impactful visuals and tweets that emphasise trust, science, and public health.

What we learned

• The importance of real-time analysis in combating misinformation.
• How to refine machine learning models to handle the nuances of social media language.
• The value of clear, empathetic communication when addressing contentious topics.
• Insights into the power of data visualisation in driving engagement and understanding.

What's next for VaxLens: Data-Driven Vaccine Insights

• Improved real-time monitoring: enhance scalability and processing speed to handle larger datasets and more platforms beyond bluesky.
• Multilingual support: expand capabilities to analyse and respond to vaccine-related conversations in multiple languages.
• Advanced analytics: incorporate deeper trend analysis, such as network mapping of misinformation sources and influence patterns.

Built With

• huggingface
• javascript
• openai
• python
• streamlit
• tavily
• uagents