Question

我正在尝试创建一个水平图来说明过程的持续时间。这是我的示例数据：

放入Jupyter Notebook的一些代码：

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as dt
import seaborn as sns

df = pd.DataFrame(
    {
    'PROC_NAME': ['data_load', 'data_send', 'data_load', 'data_send', 'data_load', 'data_send', 'data_load', 'data_send'],
    'START_TS': ['2019-06-25 03:30', '2019-06-25 07:15', '2019-06-26 03:30', '2019-06-26 07:19', 
                 '2019-06-26 08:54', '2019-06-27 03:30', '2019-06-27 08:51', '2019-06-28 03:30'],
    'END_TS': ['2019-06-25 03:51', '2019-06-25 07:52', '2019-06-26 03:40', '2019-06-26 07:43', 
               '2019-06-26 09:21', '2019-06-27 04:16', '2019-06-27 09:32', '2019-06-28 04:02']    
    })

df.head()

我想创建一个水平条形图来说明每天的运行时间，例如：

[右]

所以它应该有点像甘特图，但每个过程只有一行，一行中有多个条。甘特图会将每个实例放在单独的行中-这不是我想要实现的目标：

[错误]

感谢您的帮助。

Answer 1

知道了！非常感谢@jdhao为此answer所做的一切。（来吧，检查一下并投票！）

这又是源数据的代码-我添加了一些数据来改进示例：

Id  | PROC_NAME         | START_TS              | END_TS
---------------------------------------------------------------------
0   | data_load         | 2019-06-25 03:30:00   | 2019-06-25 03:51:00
1   | data_send         | 2019-06-25 07:15:00   | 2019-06-25 07:52:00
2   | data_load         | 2019-06-26 03:30:00   | 2019-06-26 03:40:00
3   | data_send         | 2019-06-26 07:19:00   | 2019-06-26 07:43:00
4   | data_load         | 2019-06-26 08:54:00   | 2019-06-26 09:21:00
5   | data_send         | 2019-06-27 03:30:00   | 2019-06-27 04:16:00
6   | data_load         | 2019-06-27 08:51:00   | 2019-06-27 09:32:00
7   | data_send         | 2019-06-28 03:30:00   | 2019-06-28 04:02:00
8   | data_extraction   | 2019-06-25 03:21:00   | 2019-06-25 03:51:00
9   | data_extraction   | 2019-06-25 06:45:00   | 2019-06-25 07:32:00
10  | data_extraction   | 2019-06-26 03:30:00   | 2019-06-26 06:40:00
11  | data_extraction   | 2019-06-26 07:19:00   | 2019-06-26 07:43:00
12  | data_extraction   | 2019-06-26 10:54:00   | 2019-06-26 11:21:00
13  | data_extraction   | 2019-06-27 05:30:00   | 2019-06-27 08:16:00
14  | data_extraction   | 2019-06-27 09:51:00   | 2019-06-27 11:32:00
15  | data_extraction   | 2019-06-28 02:30:00   | 2019-06-28 04:02:00

这是Jupyter的代码：

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as dt


df = pd.DataFrame(
    {
    'PROC_NAME': ['data_load', 'data_send', 'data_load', 'data_send', 'data_load', 'data_send', 'data_load', 'data_send',
                  'data_extraction', 'data_extraction', 'data_extraction', 'data_extraction', 'data_extraction', 'data_extraction', 'data_extraction', 'data_extraction',],
    'START_TS': ['2019-06-25 03:30', '2019-06-25 07:15', '2019-06-26 03:30', '2019-06-26 07:19', 
                 '2019-06-26 08:54', '2019-06-27 03:30', '2019-06-27 08:51', '2019-06-28 03:30',
                 '2019-06-25 03:21', '2019-06-25 06:45', '2019-06-26 03:30', '2019-06-26 07:19', 
                 '2019-06-26 10:54', '2019-06-27 05:30', '2019-06-27 09:51', '2019-06-28 02:30'],
    'END_TS': ['2019-06-25 03:51', '2019-06-25 07:52', '2019-06-26 03:40', '2019-06-26 07:43', 
               '2019-06-26 09:21', '2019-06-27 04:16', '2019-06-27 09:32', '2019-06-28 04:02',
               '2019-06-25 03:51', '2019-06-25 07:32', '2019-06-26 06:40', '2019-06-26 07:43', 
               '2019-06-26 11:21', '2019-06-27 08:16', '2019-06-27 11:32', '2019-06-28 04:02']  
    })

#convert input to datetime:
df.START_TS = pd.to_datetime(df.START_TS, format = '%Y-%m-%d %H:%M')
df.END_TS = pd.to_datetime(df.END_TS, format = '%Y-%m-%d %H:%M')
df.head()

使用pyplot.hlines解决我的问题：

fig = plt.figure()
fig.set_figheight(2)
fig.set_figwidth(15)
ax = fig.add_subplot(211)

plt.xticks(rotation='25')

#format dates on x axis
ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %d %H:%M'))
ax = ax.xaxis_date()
ax = plt.hlines(df.PROC_NAME,
                dt.date2num(df.START_TS),
                dt.date2num(df.END_TS),
                lw = 10, # make the lines wider and looking more like ribbon
                color = 'b' # add some color
               )

最后是结果，我可以清楚地看到运行时间和重叠部分：

用熊猫绘制水平持续时间

1 个答案: