PCA and HCA

This dataset contains the data, R script, and graphical outputs used for the multivariate statistical analysis of commercial surface-mounted device (SMD) LED diodes recovered from end-of-life tubular LED lamps. The repository accompanies the study "Magnetic separation reproduces the compositional groups of end-of-life SMD LEDs" and provides all files required to reproduce the Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA) presented in the manuscript. The analyses were performed using standardized elemental composition data obtained from representative LED samples. The R script performs data standardization, principal component analysis, calculation of explained variance, generation of score and loading plots, and hierarchical clustering using Euclidean distance and Ward's linkage method. Repository contents LED_PCA_HCA.txt – Input dataset containing the elemental composition used for the multivariate analyses. LED_PCA_HCA.R – Fully reproducible R script used to standardize the data, calculate PCA, generate score and loading plots, and perform hierarchical cluster analysis (Ward's method). (a) Scores.png – PCA score plot showing the distribution of the LED samples according to the first two principal components. (b) Loadings.png – PCA loading plot illustrating the contribution of each chemical element to the principal components. (c) HCA.png – Hierarchical clustering dendrogram (Ward's method) showing the similarity relationships among the analyzed LED samples. The first two principal components explain 75.8% of the total variance (PC1 = 59.5%, PC2 = 16.3%), allowing the discrimination of three compositional groups corresponding to Fe-rich, Cu–Zn, and Cu-rich LED substrates. The HCA independently confirms the same clustering pattern observed in the PCA score plot. These files are provided to ensure complete reproducibility of the multivariate statistical analyses and may be reused for comparison with other electronic waste datasets, methodological benchmarking, or educational purposes. If these data are used in future work, please cite both this dataset and the associated publication.