plot many signals together in python

Question

I am trying to plot many signals (196) from a multielectrode dataset (https://zenodo.org/records/1411883). I want to plot something cool to see all registers together. I want to get something similar to this:

Is there any python package or way that you would advice me to follow for plotting all this intormation?

I tried createing subplots but they are too small and take too much to generate.

Answer 1

I think plotting 196 signals on the same plot is a bit too much.

Below is for 30 signals, using a computed offset:

EDIT: Added splitting across subplots and MultiCursor.

import matplotlib.pyplot as plt
import numpy as np
from matplotlib.widgets import MultiCursor


def main():
    n, t = 196, np.linspace(0, 10, 600)
    signals = create_signals(t, n) + create_signals(t, n) + create_signals(t, n)

    nmax_per_col = 30
    n_cols = n // nmax_per_col + bool(n % nmax_per_col)

    # Split signals into groups of almost same size
    signals_groups = np.array_split(signals, n_cols, axis=0)

    fig, axes = plt.subplots(ncols=n_cols, sharex=True, figsize=(2.5 * n_cols, 8))
    for ax, signals_group in zip(axes, signals_groups):
        plot_multisignals(t, signals_group, ax=ax)

    # Need to keep "multi" alive
    multi = MultiCursor(None, tuple(axes.flat), color="r", lw=1)

    plt.show()


def plot_multisignals(t, signals, horizontal_lines=True, ax=None, **kwargs):
    """
    t: shape (N,)
    signals: shape (M, N)
    """
    if ax is None:
        _, ax = plt.subplots()

    # Center the signals and add offsets
    sigs = signals - signals.mean(axis=1, keepdims=True)
    offset = np.max(sigs.max(axis=1)[:-1] - sigs.min(axis=1)[1:])
    sigs = sigs.T + np.arange(len(signals)) * offset

    ax.plot(t, sigs, **kwargs)
    if horizontal_lines:
        for i in range(len(signals) + 1):
            ax.axhline(i * offset - offset / 2, color="k", lw=0.5, alpha=0.3)

    # Remove y-axis labels because no meaning
    ax.tick_params(left=False, labelleft=False)
    return ax


def create_signals(t, n):
    # Dummy function to generate signals
    freqs = np.random.rand(n) * 3 + 1
    amps = np.random.rand(n) * 3.8 + 0.2
    signals = np.zeros((n, len(t)))
    for i, (freq, amp) in enumerate(zip(freqs, amps)):
        signals[i] = amp * np.sin(2 * np.pi * freq * t)
    return signals


if __name__ == "__main__":
    main()