如何指定xtick在一个特定的范围

发表于 | 更新于 | 分类于

在做EDA的时候，会有种情况，时序数据所做的plot的x-axis会存在过于密集的情况，如下

所以我们要做的就是只用将数据所在范围按一定频率切分开就行了，如这个Q，[how to change the xticks to a specific range [duplicate\]](https://stackoverflow.com/questions/56713197/how-to-change-the-xticks-to-a-specific-range) 和 Changing the “tick frequency” on x or y axis in matplotlib?

分别引用自StackOverflow用户 Heike和Alexandre B.并作出了一定修改。

在comment和answer处提供了两种解决方案：

  • 添加一行代码plt.xticks(range(9, 40, 10), range(10, 41, 10))

    这里第一个参数是是将xtick分块显示，第二个参数是各个分块处显示的数字.(这里原来的数据找不到了，就用了answer中随机生成的数据为例子).

    例子:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    # Your data to count
    y = np.random.randint(0,41,1000)

    # Create plot
    fig, ax = plt.subplots()
    sns.countplot(y)

    # Show graph
    plt.xticks(range(0, 41, 10), range(0, 41, 10))
    plt.xlabel('user_id')
    plt.show()

    但是假如将第一个参数设置的分块范围超过数据已有的范围，则会在plot超出数据已有范围处显示出一段blank。

  • One way is to define the labels on the x-axis. The set_xticklabels method from matplotlibmodule do the job (doc). By defining your own labels, you can hide them by setting the label equal to ''.

    By defining your own labels, you need to take care that they are still consistent with your data.

    Here is one example:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    # import modules
    import numpy as np
    import seaborn as sns
    import matplotlib.pyplot as plt

    #Init seaborn
    sns.set()

    # Your data to count
    y = np.random.randint(0,41,1000)

    # Create the new x-axis labels 
    x_labels = ['' if i%10 != 0 else str(i) for i in range(len(np.unique(y)))]
    print(x_labels)
    # ['0', '', '', '', '', '', '', '', '', '', 
    # '10', '', '', '', '', '', '', '', '', '', 
    # '20', '', '', '', '', '', '', '', '', '', 
    # '30', '', '', '', '', '', '', '', '', '', '40']

    # Create plot
    fig, ax = plt.subplots()
    sns.countplot(y)

    # Set the new x axis labels
    ax.set_xticklabels(x_labels)
    # Show graph
    plt.show()

Reference