Despite decades of prevention, tobacco addiction is still a widespread health concern responsible for around 7 million death per year. Research shows that behavioral support through group support and personal information is an effective intervention to help people quit especially when they present tailored content. Although such technological solutions (e.g., forums) exist, they are mostly general and lack tailored feedback (e.g., personal progress reports, comparison of motivation). The goal of this project is to devise a technological solution that provides such personalized support through an innovative social media ambient visualization system called Smokwit. This system will provide an overview of the status of the community along several dimensions. However it goes beyond simple activity metrics and includes processing of unstructured data to infer emotional and motivational aspects (e.g., motivation to quit, need of support). The challenge is therefore to provide an analytics pipeline that takes in unstructured social media data and processes them through different transformation modules including machine learning in order to vizualise the several key dimensions. We will evaluate the ambient visualization system with Reddit data and users, and field experts. We consider that this research project can significantly improve support behavioral changes. 

• Date début : 01.03.2020
• Date fin : 28.02.2021
• Montant: 49 912 CHF
• Financement : projet de recherche Hasler