Parallel Coordinates Plot: A Visual Examination of Data Structures in Exploratory Data Analysis

This study introduces parallel coordinates plots as an innovative visualization method, offering an alternative to traditional score plots in Principal Component Analysis (PCA) for exploratory data analysis (EDA). EDA is pivotal for unraveling intricate data structures, and PCA serves as a powerful tool for simplifying complex datasets. The paper explores the application of parallel coordinates plots in visualizing and interpreting PCA results, emphasizing their effectiveness in revealing multivariate patterns and detecting outliers. Through the examination of simulated datasets, where two distinct classes are generated in three-dimensional and five-dimensional spaces, and real Raman spectroscopic data, the research demonstrates the utility of parallel coordinates plots in enhancing data exploration and providing valuable insights in pattern recognition methodologies. The results highlight the complementary nature of traditional score plots and parallel coordinates plots, offering a comprehensive approach to understanding high-dimensional datasets. This innovative visualization technique proves valuable in discerning subtle patterns and relationships within complex data structures, contributing to a more profound exploration of datasets in various domains.