A Methodological Framework for AI-Driven Textual Data Analysis in Digital Media
Douglas Farias Cordeiro, Carlos Lopezosa, Javier Guallar
Abstract
The growing volume of textual data generated on digital media platforms presents significant challenges for the analysis and interpretation of information. This article proposes a methodological approach that combines artificial intelligence (AI) techniques and statistical methods to explore and analyze textual data from digital media. The framework, titled DAFIM (Data Analysis Framework for Information and Media), includes strategies for data collection through APIs and web scraping, textual data processing, and data enrichment using AI solutions, including named entity recognition (people, locations, objects, and brands) and the detection of clickbait in news. Sentiment analysis and text clustering techniques are integrated to support content analysis. The potential applications of this methodology include social networks, news aggregators, news portals, and newsletters, offering a robust framework for studying digital data and supporting informed decision-making. The proposed framework is validated through a case study involving data extracted from the Google News aggregation platform, focusing on the Israel–Lebanon conflict. This demonstrates the framework’s capability to uncover narrative patterns, content trends, and clickbait detection while also highlighting its advantages and limitations.