Youtube emotional database: How to acquire user feedback to build a database of emotional video stimuli

Feedbacks are an important tool for interacting within a social environment where our daily actions are continuously dictated and influenced by others as well as, of course, our actions influence others´ choices. In this paper we propose a software architecture that offers a non-invasive and programmable environment to get user feedbacks during the navigation within the World Wide Web, and a research agenda on how to exploit the proposed tool for investigating on human social behaviour during co-present and virtual emotional interactional exchanges. The proposed software architecture comprises two components: a browser add-on, that catches direct or indirect user actions and a centralized database server that collects and aggregates data. We instantiated and tested our architecture by developing an application for collecting user emotional feedbacks during the visualization of videos on YouTube. The developed application is devoted to build up a large assessed database of emotional video stimuli labeled and rated by naive judges (the WWW users).