6.4 heatmap/imshow

# sphinx_gallery_thumbnail_number = 11

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

from easy_mpl import imshow
from easy_mpl.utils import version_info, despine_axes

version_info()  # print version information of all the packages being used

{'easy_mpl': '0.21.4', 'matplotlib': '3.8.2', 'numpy': '1.26.3', 'pandas': '1.5.3', 'scipy': '1.11.4'}

x = np.random.random((10, 8))

_ = imshow(x)

_ = imshow(x, annotate=True)

show colorbar

_ = imshow(x, colorbar=True)

do not show border around colorbar

_ = imshow(x, colorbar=True, cbar_params={"border": False})

Move the colorbar below the heatmap

_ = imshow(x, colorbar=True, cbar_params={"border": False, 'pad': 0.4,
                                          "orientation":"horizontal"})

show white grid line

data = np.random.random((4, 10))

_ = imshow(data, cmap="YlGn",
       xticklabels=[f"Feature {i}" for i in range(data.shape[1])],
       grid_params={'border': True, 'color': 'w', 'linewidth': 2},
           annotate=True,
       colorbar=True)

we can specify color of text in each box of imshow for annotation For this, textcolors must a numpy array of shape same as that of data. Each value in this numpy array will define color for corresponding box annotation.

data = np.arange(9).reshape((3,3))

_ = imshow(data, cmap="Blues",
       annotate=True,
       annotate_kws={
              "textcolors": np.array([['black', 'black', 'black'],
                                      ['black', 'black', 'black'],
                                     ['white', 'white', 'white']]),
              'fontsize': 14
       },
       colorbar=True)

We can decide which portion of heatmap to show using mask argument

x = np.random.random((20, 20))
_ = imshow(x, mask=True)

_ = imshow(x, mask="upper")

_ = imshow(x, mask="lower")

The imshow function returns Axesimage object of matplotlib which can be used for further processing. The Axesimage is not axes, but we can get the axes from Axesimage using Axesimage.axes the process it as shown below.

data = pd.read_json('https://climatereanalyzer.org/clim/t2_daily/json_cfsr/cfsr_world_t2_day.json')
index = data.pop('name')
nyrs = 45
data = pd.DataFrame(
    np.array([np.array(data.iloc[row, :].values[0]) for row in range(nyrs)]),
    index=pd.to_datetime(index[0:nyrs])
)
data = data.astype(float)
data1 = pd.concat([data.iloc[i, :] for i in range(data.shape[0])]).dropna()
data1.index = pd.date_range(data.index[0], periods=len(data1), freq="D")
mon_data = data1.resample('M').mean()

data_np = np.full(shape=(12, nyrs), fill_value=np.nan)
for ii, i in enumerate(range(0, len(mon_data), 12)):
    data_np[:, ii] = mon_data.iloc[i:i + 12].values

print(data_np.shape)

im = imshow(
    data_np,
    cmap="RdBu_r",
    aspect="auto",
    colorbar=True,
    cbar_params=dict(border=False, title="Mean Temperature",
                     title_kws=dict(fontsize=14)),
    show=False,
    ax_kws=dict(xlabel="Years", ylabel="Months",
                xlabel_kws=dict(fontsize=14), ylabel_kws=dict(fontsize=14)),
    grid_params={'border': True, 'color': 'w', 'linewidth': 0.5},
)
im.axes.set_yticks(range(12))
im.axes.set_yticklabels(
    ['Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec'])
im.axes.set_xticks(np.linspace(0, data_np.shape[-1], 6))
im.axes.set_xticklabels(np.linspace(data.index.year.min(), data.index.year.max(), 6, dtype=int))
despine_axes(im.axes)
im.axes.tick_params(axis=u'y', which=u'both',length=0)
ticklabels = []
for ticklabel in im.colorbar.ax.get_yticklabels():
    ticklabel.set_text(f"{ticklabel.get_text()}℃")
    ticklabels.append(ticklabel)
im.colorbar.set_ticklabels(ticklabels)
plt.tight_layout()
plt.show()

(12, 45)
/home/docs/checkouts/readthedocs.org/user_builds/python-seekho/checkouts/latest/scripts/plotting/_imshow.py:131: UserWarning: set_ticklabels() should only be used with a fixed number of ticks, i.e. after set_ticks() or using a FixedLocator.
  im.colorbar.set_ticklabels(ticklabels)

We can pass any valid matplotlib cmap. For example, we can use cmaps from seaborn library.

import seaborn as sns

cm = sns.color_palette("rocket_r", as_cmap=True)

print(type(cm))

<class 'matplotlib.colors.ListedColormap'>

im = imshow(
    data_np,
    cmap=cm,
    aspect="auto",
    colorbar=True,
    cbar_params=dict(border=False, title="Mean Temperature",
                     title_kws=dict(fontsize=14)),
    show=False,
    ax_kws=dict(xlabel="Years", ylabel="Months",
                xlabel_kws=dict(fontsize=14), ylabel_kws=dict(fontsize=14)),
    grid_params={'border': True, 'color': 'w', 'linewidth': 0.5},
)
im.axes.set_yticks(range(12))
im.axes.set_yticklabels(
    ['Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec'])
im.axes.set_xticks(np.linspace(0, data_np.shape[-1], 6))
im.axes.set_xticklabels(np.linspace(data.index.year.min(), data.index.year.max(), 6, dtype=int))
despine_axes(im.axes)
im.axes.tick_params(axis=u'y', which=u'both',length=0)
ticklabels = []
for ticklabel in im.colorbar.ax.get_yticklabels():
    ticklabel.set_text(f"{ticklabel.get_text()}℃")
    ticklabels.append(ticklabel)
im.colorbar.set_ticklabels(ticklabels)
plt.tight_layout()
plt.show()

/home/docs/checkouts/readthedocs.org/user_builds/python-seekho/checkouts/latest/scripts/plotting/_imshow.py:171: UserWarning: set_ticklabels() should only be used with a fixed number of ticks, i.e. after set_ticks() or using a FixedLocator.
  im.colorbar.set_ticklabels(ticklabels)