Presentation information

Organized Session

Organized Session » [OS] OS-1

[2D4-OS-1a] 計算社会科学(1)

Wed. Jun 5, 2019 3:20 PM - 5:00 PM Room D (301B Medium meeting room)

鳥海 不二夫(東京大学)、笹原 和俊(名古屋大学)、吉田 光男(豊橋技術科学大学)、瀧川 裕貴(東北大学)、榊 剛史(ホットリンク)、高野 雅典(サイバーエージェント)

3:20 PM - 3:40 PM

[2D4-OS-1a-01] Proposal of the Method to Construct Landscape Map for Massive Information Propagation on Social Media

〇Takeshi Sakaki1,2, Fujio Toriumi2, Masanao Ochi2 (1. Hottolink, Inc., 2. Univ. of Tokyo)

Keywords:Computational Social Science, Web Mining

In recent years, a lot of people pay attention to negative aspects of information publishing by individuals, such as fake news, flames and echo-chamber phenomena. However, information propagation on social media has become a large-scale and complicated phenomenon, and it is difficult for human beings to intuitively understand the whole picture.
Therefore, we propose a method to visualize information propagation on social media in an intuitively understandable form in this paper. Specifically, in the proposed method a clustering method is applied recursively to a social network on Twitter and a hierarchical structure of user communities is constructed. We use the hierarchical structure to visualize information propagation. By using a hierarchical structure, we transform complex phenomena into simpler structures, thereby encouraging human intuitive understanding. We applied the proposed method to the actual case of large information propagation on Twitter and confirmed that reasonable visualization result can be obtained.