Python矩阵作图库matplotlib的初级使用(2)

• 基础介绍
  matplotlib图形对象层级结构:
  图形对象(figure) → 子图对象(axes) → 坐标轴对象(axis) → 定位器对象-刻度线(locator)/格式化器对象-刻度线标签(formatter)

• 绘图对象创建

  from matplotlib import pyplot as plt
  
  # 创建绘图对象
  fig = plt.figure()
  
  # 创建网格子图
  ax1 = fig.add_subplot(rows, cols, idx)
  # 创建手动子图
  ax1 = fig.add_axes([left, bottom, width, height])
  
  # 获取坐标轴对象
  ax1.xaxis
  ax1.yaxis
  
  # 设置定位器对象
  ax1.xaxis.set_major_locator(plt.NullLocator)
  ax1.xaxis.set_minor_locator(plt.NullLocator)
  # 设置格式化器对象
  ax1.xaxis.set_major_formatter(plt.NullFormatter)
  ax1.xaxis.set_minor_formatter(plt.NullFormatter)
  

• 部分细节控制

  • 颜色支持
    ①. 标准颜色名称, 如: "red"
    ②. 范围在0～1的灰度值, 如: "0.75"
    ③. RGB十六进制, 如: "#FFDD44"
    ④. RGB元组, 范围在0～1, 如: (1.0, 0.2, 0.3)

  • 配色方案支持
    ①. 获取配色方案: plt.cm
    ②. 常见配色方案: "jet", "viridis", "RdBu"
    ③. 离散化配色方案: plt.cm.get_cmap("Blues", 6)

  • 颜色条控制支持
    plt.colorbar(mappable, ticks=range(6), label="digit value", extend="both")
    注意, colorbar本身也是一个子图对象, 创建时依赖于mappable对象

  • linestyle线条风格支持
    ①. "solid"; ②. "dashed"; ③. "dashdot"; ④. "dotted"

  • 散点控制支持
    ①. markersize; ②. markerfacecolor; ③. markeredgecolor; ④. markeredgewidth
    注意, 绘制散点时, plt.plot较plt.scatter性能更好, plt.scatter较plt.plot更加灵活

  • 标签支持
    ①. title; ②. xlabel; ③. ylabel; ④. label; ⑤. legend

  • 图例控制支持
    ax1.legend([labels,] loc="upper left", frameon=False, ncol=2, fontsize=10, title="Area")

  • 文本注释支持
    ax1.text(x, y, text, size, color, ha, va, transform=ax1.transData|ax1.transAxes|fig.transFigure)

  • 坐标轴控制支持
    ax1.set(xlabel, xlim, xticks, xscale)

• 部分绘图实例

  • 误差绘制

    code

    import numpy
    from matplotlib import pyplot as plt
    
    X = numpy.linspace(-1, 1, 30)
    Y = X ** 3 + X ** 2 + X
    YErr = numpy.random.uniform(0.3, 0.7, Y.shape)
    
    fig = plt.figure(figsize=(4, 4))
    ax1 = fig.add_subplot(2, 1, 1)
    ax2 = fig.add_subplot(2, 1, 2)
    
    ax1.errorbar(X, Y, YErr, fmt="o", color="black", ecolor="lightgray", elinewidth=3, capsize=0)
    ax2.plot(X, Y, "-", color="gray")
    ax2.fill_between(X, Y-YErr, Y+YErr, color="gray", alpha=0.2)
    
    fig.savefig("plot.png", dpi=300)
    

    

  • 等高线绘制

    code

    import numpy
    from matplotlib import pyplot as plt
    
    X = numpy.linspace(-1, 1, 30)
    Y = numpy.linspace(-2, 2, 50)
    X, Y = numpy.meshgrid(X, Y)
    Z = numpy.sin(X) + numpy.cos(Y ** 2)
    
    fig = plt.figure(figsize=(10, 3))
    ax1 = fig.add_axes([0.1, 0.1, 0.3, 0.9])
    ax2 = fig.add_axes([0.5, 0.1, 0.3, 0.9])
    
    ax1.contourf(X, Y, Z, 20, cmap="RdGy")
    ax2.imshow(Z, extent=[-1, 1, -2, 2], origin="lower", cmap="RdGy", aspect="auto")
    
    fig.savefig("plot.png", dpi=300)
    

    

  • 3维曲面绘制

    code

    import numpy
    from matplotlib import pyplot as plt
    from mpl_toolkits import mplot3d
    
    X = numpy.linspace(-1, 1, 30)
    Y = numpy.linspace(-2, 2, 50)
    X, Y = numpy.meshgrid(X, Y)
    Z = numpy.sin(X) + numpy.cos(Y ** 2)
    
    fig = plt.figure(figsize=(10, 3))
    ax1 = fig.add_subplot(1, 2, 1, projection="3d")
    ax2 = fig.add_subplot(1, 2, 2, projection="3d")
    
    ax1.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap="viridis", edgecolor="none")
    ax2.plot_trisurf(X.flatten(), Y.flatten(), Z.flatten(), cmap="viridis", edgecolor="none")
    
    fig.tight_layout()
    fig.savefig("plot.png", dpi=300)
    

    

  • 直方图与密度图

    code

    import numpy
    import seaborn as sns
    from matplotlib import pyplot as plt
    
    X = numpy.random.normal(0, 1, (10000,))
    
    fig = plt.figure(figsize=(10, 3))
    ax1 = fig.add_subplot(1, 2, 1)
    ax2 = fig.add_subplot(1, 2, 2)
    
    sns.histplot(X, kde=False, ax=ax1)
    sns.kdeplot(data=X, ax=ax2)
    
    fig.savefig("plot.png", dpi=300)
    

    

  • 多维矩阵图

    code

    import numpy
    import pandas
    import seaborn as sns
    
    X1 = numpy.random.normal(0, 1, (300,))
    X2 = numpy.sin(X1)
    X3 = X1 ** 2
    data = numpy.vstack((X1, X2, X3)).T
    df = pandas.DataFrame(data, columns=["X1", "X2", "X3"])
    
    sns_plot = sns.pairplot(df)
    
    sns_plot.savefig("plot.png", dpi=300)
    

    

• 参考文档
  ①. Python Data Science Handbook by Jake VanderPlas (O’Reilly). Copyright 2017 Jake VanderPlas, 978-1-491-91205-8