ワークショップ (国内) Capturing spatial distribution of people interests with web queries and location data: A large scale empirical study of metropolises in Japan

Jinyuan Li (Tokyo Institute of Technology), Soto Anno (Tokyo Institute of Technology), Takayuki Sakamoto (Nikken Sekkei Ltd.), Hiraki Yasuda (Nikken Sekkei Ltd.), Akira Tsujimoto (Nikken Sekkei Ltd.), Kota Tsubouchi, Masamichi Shimosaka (Tokyo Institute of Technology)

情報処理学会 第71回ユビキタスコンピューティングシステム (UBI) 研究発表会 (IPSJ SIGUBI)


Capturing visitor interests in certain urban areas provides a good approach to mine interests spatial distribution that is essential for more intelligent urban designing and city planning. To learn better about a city, currently it takes urban researchers much cost on field investigation such as gathering pictures or crowd-sourcing. Meanwhile, evaluation obtained from visitors is often biased and fail to fulfill our purpose to characterize interests distribution objectively. Under the help of massive query and location data gathered from users, we conduct several empirical studies on geographic map segments. In our work, spatial interests distribution can be captured by calculating statistical features across the regions. We apply the aggregation and analysis procedure on real world dataset collected by Yahoo corporation in several cities in Japan. Our analysis results can provide a new perspective on how to characterize a city without extensive field studies.