Visual Analytics for Sustainability and Climate Change/Mengyuan Fu and Kunhao Yang

Interview with Mengyuan Fu and Kunhao Yang, 5 February 2026, CC BY 4.0

OSF Open Science Framework: (to be added)
ORCID Mengyuan Fu: https://orcid.org/0000-0002-4734-1035
ORCID Kunhao Yang: https://orcid.org/0000-0002-7654-9441
Watch the unedited interview on BBB

Mengyuan Fu is a science communication researcher based in Tokyo, where she is a Junior Researcher (Assistant Professor) at Waseda University's Institute of Political Economy after completing her PhD at the University of Tokyo. She was also a visiting scholar at the University of Washington's Department of Communication. Her research examines how scientific knowledge is communicated in digital environments, including public discourse, climate misinformation, and AI-mediated information. Together with Kunhao Yang, she has published influential work using Wikipedia as a large-scale dataset to study collaborative knowledge production on climate change, and more recently has expanded this work to multilingual online knowledge and AI systems.

Kunhao Yang is a data scientist and computational social scientist based in Tokyo, currently serving as an Assistant Professor at Shibaura Institute of Technology. He previously held positions at Yamaguchi University and Waseda University, and was a visiting scholar at the University of Washington. His research combines machine learning, network science, and computational social science to study collective behaviour and online knowledge systems. A significant strand of his work uses Wikipedia as a computational laboratory to analyse collaboration, conflict, and knowledge production, alongside broader research on online communities, misinformation, and multilingual AI.

Scientific paper relevant to the project: Yang, K., Fu, M. Polarized collaboration benefits knowledge production: empirical analyses of the mediating effect of co-production pattern in Wikipedia articles on climate change. J Comput Soc Sc 7, 2677–2699 (2024). https://doi.org/10.1007/s42001-024-00321-3

Notes from the interview

notes to be edited

Mengyuan Fu (science communication) and Kunhao Yang (data science), 5 February 2026. Based in Tokyo. They studies the Fukushima Nuclear Power Plant Disaster that occurred in March 2011 in Japan on Wikipedia.

Mengyuan Fu. It was part of my PhD. Work in the Japanise contect. Published in new media Differences in different languages. Japanese compared with English. A lot of information gaps. Different development, outdated information, often in Japanese are translations from English. In particular for scientific information.

Kunhao Yang. We looked at the climate change Knowledge cop-production patterns on Wikipedia. Computetional social science - data. Collaboration patterns. We looked at how people collaborate to edit content. Identiy left wing or right wing. How editors work together.

Relevant parameters.

Mengyuan Fu. Looking at the quality of the article. * References quoted in the article

I compare references in different languages
In Wikipedia in English there are many scientific papers. In Japanese the online media are more quoted.

Kunhao Yang. Which part of the article have more edits. The distribution of the article is not even. Some controversial part can have more edit. Evaluation of the quality of the articles internal to the Wikimedia communities.

Important to look at which part of the article has the best and worst quality.

Computetional method. Editing history to identify tendencies left and right wing. Collaboration patterns. Seldom collaboration with eachother; normally they work within their group. The cross collaboration significantly contribute to the quality of the article. The relation is confirmed at a statistical level. Limited percentage.

Editing war on Wikipedia. We don't have conclusions. Some was can improve the article quality. There can be some productive wars. Encouraging editors to redefine articles and work on its content. We don't know if it is a general pattern. We want to identify which factors triggers those wars.

Translation - sometimes they just translate from English

Understanding when it is only a translation from one language to the other. Data source

Climate change articles in Japanese have only 50 source of information. The English version defined the content on the Japanese Wikipedia. Dominance of English language on the global scale.

[Iolanda] action to strengthen relationship between academia and wikipedia community they could join

japanese community unique

we have a special focus on extreme weather events/disasters - Fukushima being one of them

project: they could join if interested

They don't work with Wikidata. They collected data from Wikipedia.

They are doing research about conspiracy theorists

Concern to use AI. Vulnerability of LLM for information related to climate change. For the non English, the LLM can easily present misinformation related to climate change. A potential risk in using AI for articles non in English.

We used AI to understand the beheviour of contributors and analyse articles.

A set of misinformation about climate change. A set of data, extracted from sources. We produced manually misinformation (simulation). Creating a scenerio / simulation. Scientific consensus. We asked LLM to identify them. Japanese works much worst.

Allignment part. between linguistic versions. If genAI is used for translations is good but more likely to use

Follow up. To do:

   * send Cost action 
   * Send info about the UNESCO

events.

* partner project related to VIZWP

Suggested requirements

How contributors work together
Groups of contributors (left and right wing): monitor how they behave, computational pattern to identify them.
If contributors from different groups collaborate, as this brings better quality to the article.
Which parts of the article have more edits: important to look at which part of the article has the best and worst quality.
Identify edit wars
Understanding when it is only a translation from one language to the other. Data source. Showing how the article evolved over time after the translation