[人工智能]【Python-matplotlib】条形统计图

Python-matplotlib 条形统计图

效果图展示如下:
该代码可以处理多个实验多组观测值的展示，代码如下:

import matplotlib.pyplot as plt
import numpy as np
from matplotlib.pyplot import MultipleLocator

def plot_bar(experiment_name, bar_name, bar_value, error_value=None,):
    """

    Args:
        experiment_name: x_labels
        bar_name: legend name
        bar_value: list(len(experiment_name), each element contains a np.array(),
                   which contains bar value in each group
        error_value: list(len(experiment_name), each element contains a np.array(),
                   which contains error value in each group
    Returns:

    """

    # 用于正常显示中文标签
    # plt.rcParams["font.sans-serif"]=['SimHei']

    colors = ['lightsteelblue', 'cornflowerblue', 'royalblue', 'blue', 'mediumblue', 'darkblue', 'navy', 'midnightblue',
              'lavender', ]

    assert len(bar_value[0]) <= len(colors)  # if not try to add color to 'colors'

    plt.rcParams['axes.unicode_minus'] = False
    plt.style.use('seaborn')
    font = {'weight': 'normal', 'size': 20, }
    font_title = {'weight': 'normal', 'size': 28, }
    # bar width
    width = 0.2
    # groups of data
    x_bar = np.arange(len(experiment_name))
    # create figure
    plt.figure(figsize=(10, 9))

    ax = plt.subplot(111)  # 假如设置为221，则表示创建两行两列也就是4个子画板,ax为第一个子画板

    # plot bar

    bar_groups = []
    value = []
    for i in range(len(bar_value[0])):
        for j in range(len(experiment_name)):
            value.append(bar_value[j][i])
        group = ax.bar(x_bar - (len(experiment_name)-3-i)*width, copy.deepcopy(value), width=width, color=colors[i], label=bar_name[i])
        bar_groups.append(group)
        value.clear()


    # add height to each bar
    i = j = 0
    for bars in bar_groups:
        j = 0
        for rect in bars:
            x = rect.get_x()
            height = rect.get_height()
            # ax.text(x + 0.1, 1.02 * height, str(height), fontdict=font)
            # error bar
            if error_value:
                ax.errorbar(x + width / 2, height, yerr=error_value[j][i], fmt="-", ecolor="black",
                            elinewidth=1.2, capsize=2,
                            capthick=1.2)
            j += 1
        i += 1

    # 设置刻度字体大小
    plt.xticks(fontsize=15)
    plt.yticks(fontsize=18)
    # 设置x轴的刻度
    ax.set_xticks(x_bar)
    ax.set_xticklabels(experiment_name, fontdict=font)

    # 设置y轴的刻标注
    ax.set_ylabel("Episode Cost", fontdict=font_title)
    ax.set_xlabel('Experiment', fontdict=font_title)

    # 是否显示网格
    ax.grid(False)

    # 拉伸y轴
    ax.set_ylim(0, 7.5)
    # 把轴的刻度间隔设置为1，并存在变量里
    y_major_locator = MultipleLocator(2.5)
    ax.yaxis.set_major_locator(y_major_locator)

    # 设置标题
    plt.suptitle("Cost Comparison", fontsize=30, horizontalalignment='center')

    plt.subplots_adjust(left=0.11, bottom=0.1, right=0.95, top=0.93, wspace=0.1, hspace=0.2)
    # 设置边框线宽为2.0
    ax.spines['bottom'].set_linewidth('2.0')
    # 添加图例
    ax.legend(loc='upper left', frameon=True, fontsize=19.5)
    # plt.savefig("test.png")
    plt.show()
    plt.legend()

if __name__ == "__main__":
    test_experiment_name = ["Test 1", "Test 2", "Test 3", "Test 4"]
    test_bar_name = ['A', "B", "C"]
    test_bar_value = [
        np.array([1, 2, 3]),
        np.array([4, 5, 6]),
        np.array([3, 2, 4]),
        np.array([5, 2, 2])
    ]
    test_error_value = [
        np.array([1, 1, 2]),
        np.array([0.2, 0.6, 1]),
        np.array([0, 0, 0]),
        np.array([0.5, 0.2, 0.2])
    ]
    plot_bar(test_experiment_name, test_bar_name, test_bar_value, test_error_value)