Extracting locations from sport and exercise-related social media messages using a neural network-based bilingual toponym recognition model

Sport and exercise contribute to health well-being in cities. While previous research has mainly focused on activities at specific locations such as sport facilities, "informal sport" that occur arbitrary across the city have been largely neglected. Such are more challenging observe, but this challenge may be addressed using data collected from social media platforms, because users regularly generate content related sports given locations. This allows studying all sport, including those which locations, better understand exercise-related However, user-generated geographical information available platforms is becoming scarcer coarser. places increased emphasis extracting location free-form text media, complicated by multilingualism informal language. To support effort, article presents an end-to-end deep learning-based bilingual toponym recognition model for exercise. We show our approach outperforms five state-of-the-art learning machine models. further demonstrate how can deployed a geoparsing framework planners promoting healthy active lifestyles.