Detecting Relations of Hotspots using Geo-tagged Photographs in Social Media Sites
Masaharu Hirota, Motohiro Shirai, Hiroshi Ishikawa, Shohei Yokoyama · 2014
Social media sites include many photographs taken at various locations and times. A hotspot at which many photographs have been taken might be an interesting place for many people to visit. Visualization of hotspots reveals user interest, which is important for industries such as tourism and marketing research. Hotspots have four types based on reasons that people take photographs: areas of interest and shooting spots that occurred because of an event or not. In some cases, a hotspot has a relation to other hotspots. We propose a method to detect and visualize relations among hotspots to organize those hotspots. Our proposed method discovers two relation types: photographic subjects and shooting spots, and tag relations using geotagging of photographs posted on social media sites. We demonstrate our approach by discovering the relation using photographs with tags, photograph orientation, and photograph locations from Flickr.