Synthetic EIS spectra for binary supervised-learning practice

Synthetic electrochemical impedance spectroscopy (EIS) data for teaching supervised classification, model tuning, underfitting, overfitting, validation and sensitivity to noise. The CSV contains 600 independent spectra. Label 1 denotes a simplified Randles circuit, Rs + (Rct parallel Cdl), without diffusion. Label 0 denotes Rs + (R1 parallel C1) + (R2 parallel C2). Each spectrum uses the same 41 logarithmically spaced frequencies from 0.1 Hz to 100 kHz, in ascending order. Each row is one spectrum. Columns are spectrum_id, label, and Zreal_00 to Zreal_40, Zimag_00 to Zimag_40, Zmod_00 to Zmod_40, and phase_00 to phase_40. Impedance is in ohms. Phase is in degrees. Zimag is the signed imaginary impedance. Suffixes index numpy.logspace(-1, 5, 41). Use only the spectral measurements as model inputs, never spectrum_id or label. All four representations are calculated from the same noisy complex impedance. Independent Gaussian measurement errors were added to its real and imaginary components, each with standard deviation 4% of the clean impedance modulus. These are simulated teaching data, not experimental measurements.