Materials play a pivotal role in manufacturing, serving as the foundation upon which the functionality and overall quality of products are built. In the material science domain, an overwhelming amount of knowledge is generated and stored as text encoding a humongous amount of information related to materials performance along the product life cycle that results fundamental in the manufacturing landscape, addressing adaptability and circularity . This study explores the application of Natural Language Processing techniques to analyze data availability, with a specific focus on the domain of polyvinyl chloride materials across the chemical, environmental, health, social and economic dimensions. While acknowledging the expanse of available academic data, this research also ventures into exploring vast web platforms, not often emphasized in the existing literature. Latent Dirichlet Allocation is employed to autonomously extract interconnected topics from textual data, providing a flexible tool to structure multifaceted datasets. Furthermore, its integration with question-and-answer schemes, powered by Large Language Models, represents a step forward in comprehensive data mapping. This combination aids in expediting the extraction of relevant information while contributing to the creation of a structured database where all relevant information pertaining to a particular topic is organized, identifying specific missing data or noncorrelated information. This approach promises to contribute to the evolution of data analysis methodologies, offering insights into the data landscapes of material science with impact in the current manufacturing scenario.