累积标准化直方图答案

【问题标题】：Cumulative Normed Histogram累积标准化直方图
【发布时间】：2018-06-13 16:29:56
【问题描述】：

我编写了一个代码，它给出了一个累积标准化直方图。

请问如何固定X轴？

此直方图具有应用于第二个维度的阈值的附加特征，因此可以使用有关“B”列的信息以及“A”列。

它还可以调整计数标准化的数字“C”。

import pandas as pd
import numpy  as np

# Data
df_1 = pd.DataFrame({'A': [1,2,1,2,3,4,2,1,4],
                     'B': [2,1,2,1,2,3,4,2,1]})

# Cumulative Normed Histogram
bins          = np.arange(0, 5, .2)
df_1['A_Bin'] = pd.cut(df_1['A'], bins=bins)

# Apply a threshold to B
df_2          = df_1[df_1['B'] > 2]

# Get the number of rows
C             = len(df_1.index)

def fun(g):
    try:
        return float(g.shape[0]) / C
    except ZeroDivisionError:
        return np.nan

hist       = df_1.groupby('A_Bin').apply(fun)
hist_2     = df_2.groupby('A_Bin').apply(fun)

hist_cum   = hist.cumsum()
hist_2_cum = hist_2.cumsum()

hist_cum.plot()
hist_2_cum.plot()

我试过这个：

import matplotlib.pyplot as plt
plt.xticks((0,2,4,6),('0','2','4','6'))

但是得到了这个：

【问题讨论】：

标签： python histogram cumulative-sum

【解决方案1】：

我只需要获取刻度并将其放入这样的 Dataframe 中：

    import pandas            as pd
    import numpy             as np

    # Data
    df_1 = pd.DataFrame({'A': [1,2,1,2,3,4,2,1,4],
                         'B': [2,1,2,1,2,3,4,2,1]})

    # Cumulative Normed Histogram
    Min  = 0
    Max  = 6
    Step = .5

    bins          = np.arange(Min, Max, Step)
    df_1['A_Bin'] = pd.cut(df_1['A'], bins=bins)

    # Apply a threshold to B
    df_2          = df_1[df_1['B'] > 2]

    # Get the number of rows
    C             = len(df_1.index)

    def fun(g):
        try:
            return float(g.shape[0]) / C
        except ZeroDivisionError:
            return np.nan

    hist                  = df_1.groupby('A_Bin').apply(fun)
    hist_2                = df_2.groupby('A_Bin').apply(fun)

    hist_cum              = hist.cumsum()
    hist_2_cum            = hist_2.cumsum()

    # Put the Histogram in a Dataframe
    df_hist_cum           = hist_cum.to_frame()
    df_hist_2_cum         = hist_2_cum.to_frame()

    # Define the Ticks
    ticks                 = np.arange(Min, (Max-Step), Step)

    df_hist_cum['X']      = ticks
    df_hist_2_cum['X']    = ticks

    df_hist_cum.columns   = ['All', 'A']
    df_hist_2_cum.columns = ['2', 'A']

    ax = df_hist_cum.plot(x='A', y='All')
    df_hist_2_cum.plot(x='A', y='2', ax=ax)

【讨论】：