Jingle Jungle Maps - Capturing Urban Sounds and Emotions in Maps

People living in urban areas suffer from increasing noise pollution. Local and national governments invest efforts in the development of sensor-based noise monitoring systems and long-term plans to reduce noise pollution. However, deploying such sensor-based systems on a city scale is very expensive. In this work, Jingle Jungle Maps, a tool for the visualization and localization of sounds in urban spaces is proposed. By means of social media data gathered from Flickr, the researchers attempt to perform a text-based analysis on each post to identify noise sources and localize them on a map. An approximation of the emotions these sounds might evoke and their perceived level of loudness are also estimated with fuzzy logic and computing with words methods. A prototype showing the results of this analysis on ten thousand posts from the city of Bern, Switzerland, is presented in this article. Furthermore, an evaluation on five hundred posts showed acceptable precision levels for the estimation of sound sources.