Experimental dataset and reproducible source code accompanying the manuscript "Clustering recovered 18650 lithium-ion cells to improve homogeneity in second-life battery pack construction" (PONE-D-26-06286), submitted to PLOS ONE. The dataset contains electrical characterization measurements for 92 recovered 18650 lithium-ion cells extracted from end-of-life laptop batteries, including discharged capacity, charge/discharge voltage profiles, internal resistance (three replicates per cell via OPUS BT-C3400 Quick Test), manufacturer identification, and chemistry labels (LCO, NMC, NCA, LMO). Seven Python scripts reproduce all analyses reported in the manuscript: K-means and GMM clustering (k = 2–10), seed-sensitivity analysis (100 random states), weighting coefficient sensitivity (w = 0–1), measurement repeatability with Monte Carlo noise propagation, random-selection baseline (N = 10,000 packs) with statistical significance testing, and post-hoc cluster-chemistry correspondence analysis. Pre-computed supplementary tables (S1–S6) are included in Excel format.