直方图

演示如何使用matplotlib绘制直方图。

  1. import matplotlib.pyplot as plt
  2. import numpy as np
  3. from matplotlib import colors
  4. from matplotlib.ticker import PercentFormatter
  6. # Fixing random state for reproducibility
  7. np.random.seed(19680801)

生成数据并绘制简单的直方图

要生成一维直方图，我们只需要一个数字矢量。对于二维直方图，我们需要第二个矢量。我们将在下面生成两者，并显示每个向量的直方图。

  1. N_points = 100000
  2. n_bins = 20
  4. # Generate a normal distribution, center at x=0 and y=5
  5. x = np.random.randn(N_points)
  6. y = .4 * x + np.random.randn(100000) + 5
  8. fig, axs = plt.subplots(1, 2, sharey=True, tight_layout=True)
  10. # We can set the number of bins with the `bins` kwarg
  11. axs[0].hist(x, bins=n_bins)
  12. axs[1].hist(y, bins=n_bins)

直方图示例

更新直方图颜色

直方图方法(除其他外)返回一个修补程序对象。这使我们可以访问所绘制对象的特性。使用这个，我们可以根据自己的喜好编辑直方图。让我们根据每个条的y值更改其颜色。

  1. fig, axs = plt.subplots(1, 2, tight_layout=True)
  3. # N is the count in each bin, bins is the lower-limit of the bin
  4. N, bins, patches = axs[0].hist(x, bins=n_bins)
  6. # We'll color code by height, but you could use any scalar
  7. fracs = N / N.max()
  9. # we need to normalize the data to 0..1 for the full range of the colormap
  10. norm = colors.Normalize(fracs.min(), fracs.max())
  12. # Now, we'll loop through our objects and set the color of each accordingly
  13. for thisfrac, thispatch in zip(fracs, patches):
  14.     color = plt.cm.viridis(norm(thisfrac))
  15.     thispatch.set_facecolor(color)
  17. # We can also normalize our inputs by the total number of counts
  18. axs[1].hist(x, bins=n_bins, density=True)
  20. # Now we format the y-axis to display percentage
  21. axs[1].yaxis.set_major_formatter(PercentFormatter(xmax=1))

直方图示例2

绘制二维直方图

要绘制二维直方图，只需两个长度相同的向量，对应于直方图的每个轴。

  1. fig, ax = plt.subplots(tight_layout=True)
  2. hist = ax.hist2d(x, y)

直方图示例3

自定义直方图

自定义2D直方图类似于1D情况，您可以控制可视组件，如存储箱大小或颜色规格化。

  1. fig, axs = plt.subplots(3, 1, figsize=(5, 15), sharex=True, sharey=True,
  2.                         tight_layout=True)
  4. # We can increase the number of bins on each axis
  5. axs[0].hist2d(x, y, bins=40)
  7. # As well as define normalization of the colors
  8. axs[1].hist2d(x, y, bins=40, norm=colors.LogNorm())
  10. # We can also define custom numbers of bins for each axis
  11. axs[2].hist2d(x, y, bins=(80, 10), norm=colors.LogNorm())
  13. plt.show()

直方图示例4

下载这个示例