Social Science Analysis Using Social Media Data to Evaluate Public Interest in River

Abstract

 To understand public’s image of rivers and to use it for future river management, social media data were analyzed. Among data of various social media, Instagram data were chosen for their real-time nature with low noise. Data were collected through hash-tag searches with hash-tags related to rivers such as #river or #Hirose-river. Through text mining and analyses of images and the number of posts, five unique outcomes were archived. First, river interest originates from private business and events. Secondly, an observed number of river users and their purposes do not always correspond with those estimated from Instagram data. Thirdly, with the index defined as the number of posts with #river-name per basin population, touristic rivers such as Shimanto River and Niyodo River have an 8-50 times higher index than the average, which indicates that tourism contributes greatly to river interest. Fourth, compared to posts with #river and posts with #mountain, #river posts concentrated on July and August and rivers do not mention particular names of places, although mountain lovers aim at a particular mountain. Finally, playing in shallow river water areas is popular.