o. taylor plot

This file shows the usage of taylor_plot() function.

A Taylor plot can be used to show statistical summary of one or more measurements/models.

import numpy as np
from easy_mpl import taylor_plot
from easy_mpl.utils import version_info

version_info()

# sphinx_gallery_thumbnail_number = -1

{'easy_mpl': '0.21.4', 'matplotlib': '3.8.0', 'numpy': '1.26.1', 'pandas': '2.1.1', 'scipy': '1.11.3'}

# The desired covariance matrix.
cov = np.array(
       [[1,  0.8, 0.6, 0.4, 0.2],
       [0.8, 1.2, 0.8, 0.6, 0.4],
       [0.6, 0.8, 0.8, 0.8, 0.6],
       [0.4, 0.6, 0.8, 1.4, 0.8],
       [0.2, 0.4, 0.6, 0.8, 0.6]]
)

# Generate the random samples.
rng = np.random.default_rng(313)
data = rng.multivariate_normal(np.zeros(5), cov, size=100)
print(data.shape)

observations =  data[:, 0]
simulations =  {"LSTM": data[:, 1],
            "CNN": data[:, 2],
            "TCN": data[:, 3],
            "CNN-LSTM": data[:, 4]}
_ = taylor_plot(observations=observations,
            simulations=simulations,
            title="Taylor Plot")

/home/docs/checkouts/readthedocs.org/user_builds/easy-mpl/checkouts/latest/examples/taylor_plot.py:35: RuntimeWarning: covariance is not symmetric positive-semidefinite.
  data = rng.multivariate_normal(np.zeros(5), cov, size=100)
(100, 5)

multiple taylor plots in one figure

def create_data(cov, seed=313, mu=np.zeros(5), size=100):

    # Generate the random samples.
    rng = np.random.default_rng(seed)

    return rng.multivariate_normal(np.zeros(5), cov, size=size)

cov1 = np.array(
    [[1,  0.8, 0.6, 0.4, 0.2],
    [0.8, 1.2, 0.8, 0.6, 0.4],
    [0.6, 0.8, 0.8, 0.8, 0.6],
    [0.4, 0.6, 0.8, 1.4, 0.8],
    [0.2, 0.4, 0.6, 0.8, 0.6]]
)

cov2 = np.array(
    [[1,  0.8, 0.6, 0.4, 0.2],
    [0.8, 1.2, 0.8, 0.6, 0.4],
    [0.6, 0.8, 0.8, 0.8, 0.6],
    [0.4, 0.6, 0.8, 1.4, 0.8],
    [0.2, 0.4, 0.6, 0.8, 0.6]]
)

cov3 = np.array(
    [[1,  0.8, 0.6, 0.4, 0.2],
    [0.8, 1.2, 0.8, 0.6, 0.4],
    [0.6, 0.8, 0.8, 0.8, 0.6],
    [0.4, 0.6, 0.8, 1.4, 0.8],
    [0.2, 0.4, 0.6, 0.8, 0.6]]
)

cov4 = np.array(
    [[1,  0.8, 0.6, 0.4, 0.2],
    [0.8, 1.2, 0.8, 0.6, 0.4],
    [0.6, 0.8, 0.8, 0.8, 0.6],
    [0.4, 0.6, 0.8, 1.4, 0.8],
    [0.2, 0.4, 0.6, 0.8, 0.6]]
)

site1_data = create_data(cov1)
site2_data = create_data(cov2)
site3_data = create_data(cov3)
site4_data = create_data(cov4)

observations = {
    'site1': site1_data[:, 0],
    'site2': site2_data[:, 0],
    'site3': site3_data[:, 0],
    'site4': site4_data[:, 0],
}

simulations = {
    "site1": {"LSTM": site1_data[:, 1],
                "CNN": site1_data[:, 2],
                "TCN": site1_data[:, 3],
                "CNN-LSTM": site1_data[:, 4]},

    "site2": {"LSTM": site2_data[:, 1],
                "CNN": site2_data[:, 2],
                "TCN": site2_data[:, 3],
                "CNN-LSTM": site2_data[:, 4]},

    "site3": {"LSTM": site3_data[:, 1],
                "CNN": site3_data[:, 2],
                "TCN": site3_data[:, 3],
                "CNN-LSTM": site3_data[:, 4]},

    "site4": {"LSTM": site4_data[:, 1],
                "CNN": site4_data[:, 2],
                "TCN": site4_data[:, 3],
                "CNN-LSTM": site4_data[:, 4]},
}

# define positions of subplots

rects = dict(site1=221, site2=222, site3=223, site4=224)

_ = taylor_plot(observations=observations,
            simulations=simulations,
            axis_locs=rects,
            plot_bias=True,
            cont_kws={'colors': 'blue', 'linewidths': 1.0, 'linestyles': 'dotted'},
            grid_kws={'axis': 'x', 'color': 'g', 'lw': 1.0},
            title="mutiple subplots")

/home/docs/checkouts/readthedocs.org/user_builds/easy-mpl/checkouts/latest/examples/taylor_plot.py:55: RuntimeWarning: covariance is not symmetric positive-semidefinite.
  return rng.multivariate_normal(np.zeros(5), cov, size=size)

using statistics instead of arrays

observations = {'std': 3.5}
predictions = {   # pbias is optional
        'Model 1': {'std': 2.80068, 'corr_coeff': 0.49172, 'pbias': -8.85},
        'Model 2': {'std': 3.8, 'corr_coeff': 0.67, 'pbias': -19.76},
        'Model 3': {'std': 3.9, 'corr_coeff': 0.596, 'pbias': 7.81},
        'Model 4': {'std': 2.36, 'corr_coeff': 0.27, 'pbias': -22.78},
        'Model 5': {'std': 2.97, 'corr_coeff': 0.452, 'pbias': -7.99}}



_ = taylor_plot(observations,
    predictions)

with customized markers

cov = np.array(
    [[1, 0.8, 0.6, 0.4, 0.2],
     [0.8, 1.2, 0.8, 0.6, 0.4],
     [0.6, 0.8, 0.8, 0.8, 0.6],
     [0.4, 0.6, 0.8, 1.4, 0.8],
     [0.2, 0.4, 0.6, 0.8, 0.6]]
)

data = create_data(cov)

observations = data[:, 0]
simulations = {"LSTM": data[:, 1],
               "CNN": data[:, 2],
               "TCN": data[:, 3],
               "CNN-LSTM": data[:, 4]}
_ = taylor_plot(observations=observations,
                simulations=simulations,
                marker_kws={'markersize': 10, 'markeredgewidth': 1.5,
                            'markeredgecolor': 'black', 'lw': 0.0})

with customizing bbox

_ = taylor_plot(observations=observations,
            simulations=simulations,
            title="custom_legend",
            leg_kws={'facecolor': 'white',
                'edgecolor': 'black','bbox_to_anchor':(0.80, 1.1),
                'fontsize': 14, 'labelspacing': 1.0},
                marker_kws = {'ms':'20', 'markeredgecolor': 'k', 'lw': 0.0},
                )