Crowdbreaks is a health trend tracking system, currently developed by the Digital Epidemiology Lab at EPFL in Switzerland. The system is designed to track trends about health and disease-related issues in real-time across different countries. By obtaining feedback from users the system will eventually improve and be able to draw more accurate conclusions about these issues.
How does it work?
Crowdbreaks collects tweets with keywords that might be related to specific health topics. By using natural language processing and machine learning techniques the system tries to filter relevant from non-relevant content. By providing more meta information (labelling) of the tweets, these algorithms will continuously improve to detect tweets which are truly relevant and understand their content. For a more in-depth explanation of the inner workings of the platform, please refer to our latest work:
Müller, Martin M., and Marcel Salathé. "Crowdbreaks: Tracking Health Trends Using Public Social Media Data and Crowdsourcing." Frontiers in public health 7 (2019).
Millions of people use social media every day to share how they feel, both good and bad. This is, in a sense, like having millions of sensors which can help us understand the health situation in real-time. By asking people to help make sense of the reported data, we invoke the wisdom of the crowd to help us separate signal from noise. We also hope to engage the public in public health and try to show the importance of these issues.
I'm interested in the data
We're happy to share the data with you. Please find all the relevant information here.
I'm interested in the code
The platform is fully open source under MIT license. Feel free to check out our Github repo - Any contributions are welcome!
You can help us by answering the questions in the projects section. Feel free to create an account if you want to keep track of your efforts. You can also provide any other feedback by writing to us email@example.com.