EMOJI-DRIVEN VISUAL SENTIMENT INTELLIGENCE FOR LOCAL GOVERNANCE COMMUNICATION AND CITIZEN ENGAGEMENT

Abstract

The rapid expansion of social media has significantly transformed communication practices between local governments and citizens, making digital symbols such as emojis an integral part of contemporary public discourse. Emojis serve as concise, language-independent visual cues that convey emotions, opinions, and attitudes, thereby enriching textual and visual communication in online civic spaces. In the context of local self-government, understanding emoji usage in citizen-generated content is increasingly important for interpreting public sentiment, enhancing participatory governance, and improving the responsiveness of local administrations. Despite notable advances in neural network models for text-based emoji prediction, forecasting emojis directly from images shared on digital platforms remains a challenging and underexplored task, particularly within governance-oriented communication analysis. This study proposes an integrated deep learning framework that combines Convolutional Neural Network (CNN)-based image classification with emoji2vec embeddings aligned with word2vec representations to predict emoji labels from images relevant to civic and local governance contexts. In addition, sentiment analysis of associated textual content is employed to improve the accuracy of future emoji prediction and to capture emotional patterns embedded in citizen–government interactions. The proposed approach effectively models semantic and sentiment relationships among emojis, enabling faster and more reliable prediction of image-based emoji usage. By optimizing the search duration and improving interpretability, the framework offers practical value for local authorities seeking data-driven insights into public perception, digital participation trends, and community feedback. Overall, the study contributes to the growing field of computational social analysis by demonstrating how visual sentiment intelligence can support evidence-based decision-making and communication strategies in local self-government.