Analysis of relationships among variables in nationwide big data of geotechnical information in Japan

Recent advancements by national institutions in Japan have significantly enhanced the accessibility of geotechnical information, enabling researchers to utilize extensive datasets via online platforms. While these datasets have been widely employed in various studies, systematic analyses of relationships among variables within large-scale geotechnical data remain limited. This study aims to address this gap by analyzing relationships between variables using a comprehensive nationwide dataset of soil tests provided by the National Geo-Information Center (NGIC). The analysis of soil hydraulic conductivities revealed a strong dependence on the proportion of fine-grained components, such as clay and silt fractions. However, correlation analysis indicated that the strongest relationship, observed with the clay fraction, yielded a correlation coefficient of -0.51, suggesting a moderate association. Further investigation into variables such as dry density, natural water content, and void ratio demonstrated their dependence on the proportion of fine-grained fractions. Notably, the upper and lower bounds of these variables were influenced by fine particle content. A particularly significant finding was the observation that as the proportion of fine particles decreased, the void ratio also declined, leading to an increase in the permeability coefficient. These results provide valuable insights into the relationships between geotechnical properties and particle-size composition, offering a novel perspective on soil behavior. This study highlights the potential of utilizing extensive geotechnical datasets to advance our understanding of soil properties and their dependencies. The findings contribute not only to the theoretical understanding of geotechnical systems but also to practical applications in geotechnical engineering, providing a foundation for future research and data-driven approaches to soil analysis.