Dataset and analysis code for: Bioleaching of Waste Printed Circuit Boards - Mechanisms, Process Optimization, and the Emerging Role of AI/ML

Description

This dataset underlies the statistical analysis in the review "Bioleaching of Waste Printed Circuit Boards: Mechanisms, Process Optimization, and the Emerging Role of AI/ML." It comprises 1,479 biological copper-recovery datapoints extracted from 49 peer-reviewed bioleaching studies, spanning four mechanism families (chemolithotrophic, fungal, cyanogenic, and heterotrophic/enzymatic). Each row records copper recovery (%) alongside seven process variables — leaching time, initial pH, pulp density, temperature, stirring speed, particle size, and inoculum size — plus mechanism/family classification. Imputed columns (suffixed _clean) fill gaps in pulp density and other variables using a documented hierarchical imputation method; raw (non-imputed) columns are retained for transparency. analysis_corrected.py is the full, deterministic analysis pipeline used to generate every statistical result in the manuscript: mechanism-family classification, ANOVA/Kruskal-Wallis tests, Tukey HSD, Spearman correlations (pooled and study-level), pulp-density imputation, multivariate regression with cluster-robust standard errors, a complete-case sensitivity regression, PCA, and a robustness-verdict summary table. Running this script against the included CSV reproduces the manuscript's results exactly (independently verified cell-by-cell and pixel-by-pixel).

Steps to reproduce

Run python analysis_corrected.py bio_data_v2_corrected.csv <output_dir> with both files in the same directory. This regenerates the full results workbook (family classification, ANOVA/Kruskal-Wallis, correlations, cluster-robust regression, PCA, robustness summary) and five figures, reproducing the manuscript's statistical results exactly.

Institutions

• Indian Institute of Technology Kharagpur

  West Bengal, Kharagpur

Categories

Licence