Seongbum Seo

Analysis of public opinions collected from digital media helps organizations maintain positive relationships with the public. Such public relations (PR) analysis often involves assessing opinions, for example, measuring how strongly people trust an organization. Pre-trained Large Language Models (LLMs) hold great promise for supporting Organization-Public Relationship Assessment (OPRA) because they can map unstructured public text to OPRA dimensions and articulate rationales through prompting. However, adapting LLMs for PR analysis typically requires fine-tuning on large labeled datasets, which is both labor-intensive and knowledge-intensive, making it difficult for PR researchers to apply these models. In this paper, we present OPRA-Vis, a visual analytics system that leverages LLMs for OPRA without requiring extensive labeled data. Our framework employs Chain-of-Thought prompting to guide LLMs in analyzing public opinion data by incorporating PR expertise directly into the reasoning process. Furthermore, OPRA-Vis provides visualizations that reveal the clues and reasoning paths used by LLMs, enabling users to explore, critique, and refine model decisions. We demonstrate the effectiveness of OPRA-Vis through two real-world use cases and evaluate it quantitatively, through comparisons with alternative LLMs and prompting strategies, and qualitatively, through assessments of usability, effectiveness, and expert feedback.

Approaches to enhancing data quality (DQ) are classified into two main categories: data- and process-driven. However, prior research has predominantly utilized batch data preprocessing within the data-driven framework, which often proves insufficient for optimizing machine learning (ML) model performance and frequently leads to distortions in data characteristics. Existing studies have primarily focused on data preprocessing rather than genuine data quality improvement (DQI). In this paper, we introduce d-DQIVAR, a novel visual analytics system designed to facilitate DQI strategies aimed at improving ML model performance. Our system integrates visual analytics techniques that leverage both data-driven and process-driven approaches. Data-driven techniques tackle DQ issues such as imputation, outlier detection, deletion, format standardization, removal of duplicate records, and feature selection. Process-driven strategies encompass evaluating DQ and DQI procedures by considering DQ dimensions and ML model performance and applying the Kolmogorov-Smirnov test. We illustrate how our system empowers users to harness expert and domain knowledge effectively within a practical workflow through case studies, evaluations, and user studies.

The study introduces new approaches that integrate public relations concepts to identify adopters of AI technologies. It seeks to discover novel methods for identifying target groups likely to adopt AI technologies, even in contexts where these technologies are not yet familiar. The study achieves this by employing latent profile analysis, regression, and structural equation modeling using data from a South Korean online panel survey (N = 625). The results identify five distinct public profiles based on their relationship assessments, problem recognition, and general trust in AI: the Cautious, Balanced, Uninterested, Confident, and Enthusiastic. Notably, the Enthusiastic group—characterized by high trust and strong relationships with the service provider—showed the strongest interest in adopting AI, highlighting their potential as key publics for introducing new AI services in business contexts. Additionally, the article contributes to public segmentation theory through the use of the situational theory of problem-solving and enhances the applicability of established public relations frameworks to the evolving field of AI.

Seongbum Seo, Sangbong Yoo, Hyelim Lee, Yun Jang, Ji Hwan Park, Jeong-Nam Kim. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations). 2025.

Public relations (PR) aims to establish positive relationships of company-customer and government-citizen. PR researchers and scholars achieve this by analyzing communication behavioral data, which encompasses the public’s experiences in consumption, policy, and governance. However, labeling this data for analysis in PR is costly and complex, primarily due to its multi-label and the need for human judgment based on the Organization-Public Relationship Assessment (OPRA) concepts. Consequently, these problems limit the development of automated labeling models because more training data is needed for fine-tuning. In this paper, we propose a novel tag cloud visualization integrated with dimensionality reduction for conceptualizing communication behavioral data to address limitations in PR. We use attention-based embedding and dimensionality reduction to visualize the public’s experience as a scatter plot according to OPRA concepts. Additionally, we provide PR researchers and scholars with clues for multi-labeling by combining the tag cloud using Bag of Words (BoW) in a scatter plot. We evaluate our proposed interactive visualization through a case study and feedback from two PR researchers.