根据组绘制条形图

Question

我有Telco-Customer Churn dataset。在根据不同的任期持续时间分析流失率之后，我想把它想象成如下图所示，其中只有针对不同“任期”的“任期”绘制了搅动的数量。

以下是我的尝试：

import pandas as pd
import matplotlib.pyplot as plt

user_data = pd.read_csv("https://github.com/WedamN/Telco-Churn-Prediction/blob/master/CustomerChurnData.csv")

# bin the tenure into every 6 months
user_data['tenure_bin'] = pd.cut(user_data['Tenure'], list(range(0, 73, 6)))

# some basic analysis
churn_rate_according_to_tenure = user_data.groupby('tenure_bin').Churn.value_counts('Yes')*100

# plot the results
churn_rate_according_to_tenure.plot().bar()
plt.show()

这是我获得的情节（有点凌乱），其中显示了“是”和“否”类别。我怎样才能解决这个问题，我只想在条形图颜色相同的情况下只显示“是”类别？

Answer 1

如果unstack想要将两个类别放在一起，我认为您需要重塑：

print (churn_rate_according_to_tenure.unstack())
Churn              No        Yes
tenure_bin                      
(0, 6]      46.666667  53.333333
(6, 12]     64.113475  35.886525
(12, 18]    67.700730  32.299270
(18, 24]    75.420168  24.579832
(24, 30]    78.190255  21.809745
(30, 36]    78.553616  21.446384
(36, 42]    78.100264  21.899736
(42, 48]    83.812010  16.187990
(48, 54]    83.809524  16.190476
(54, 60]    87.378641  12.621359
(60, 66]    90.712743   9.287257
(66, 72]    94.703390   5.296610

churn_rate_according_to_tenure.unstack().plot.bar()

如果想要过滤，请选择类别 - 列：

churn_rate_according_to_tenure.unstack()['Yes'].plot.bar()