Work

CN-VEFD: A visual emotion feature dataset for Chinese ESG reports

Year: 2025

Type: data-paper

Abstract: As ESG reporting gains prominence, visual elements play a critical role in shaping stakeholder perceptions, yet most research prioritizes textual analysis and overlooks visual emotion. We present CN-VEFD, a large-scale visual emotion feature dataset derived from 13,481 ESG reports by Chinese listed companies spanning 2006-2023. The dataset contains 59 structured visual emotion features computed from 399,321 images, covering three dimensions: color emotion, composition emotion, and facial emotion. In addition, 7 basic information fields are provided to link images to reports and firms (total 66 fields). Features are quantified through an automated pipeline that integrates color analysis, saliency detection, and facial emotion recognition. CN-VEFD supports research in visual analytics, emotion recognition, and behavioral finance, enabling systematic exploration of how visual emotional signals in ESG disclosures relate to corporate governance and stakeholder responses.

Source: Data in Brief

DOI: 10.1016/j.dib.2025.112305

Authors: Zhao Duan, Yitong Xu, Binglong Xia

Institution: Central China Normal University

Language: English

Topic: Corporate Social Responsibility Reporting

Subfield: Strategy and Management

Field: Business, Management and Accounting

Domain: Social Sciences

Open Access status: gold

Estimated APC paid: $1,220

Funders: Major Program of National Fund of Philosophy and Social Science of China, National Social Science Fund of China