根据时间增量值生成图

Question

我有一个数据框，可以捕获从服务器发送的数据。服务器数据至少每 5 分钟一次。如果服务器超过 5 分钟没有发送数据，直到再次发送数据的时间被认为是中断。我想在图表中可视化这些停电。数据框看起来像

timestamp                   temperature
2019-06-03 14:16:31.149132  27.17
2019-06-03 14:21:34.732911  27.13
2019-06-03 14:37:20.437143  27.16
2019-06-03 14:42:15.516416  27.13
2019-06-03 14:51:26.167553  27.19
2019-06-03 14:56:31.244862  27.02
2019-06-03 15:07:30.519727  27.1
2019-06-03 15:12:57.319953  27.12
2019-06-03 15:17:56.256638  27.12

我计算了两个时间戳之间的时间差，并标记了停电并计算了停电时间。代码：

df['TimeDelta'] = df['timestamp'] - df['timestamp'].shift()
df['blackout'] = np.where(df['TimeDelta'] > datetime.timedelta(minutes = 5) , 1 , 0)
df['blackoutTime'] =  np.where(df['blackout'] > 0, df['TimeDelta'] - datetime.timedelta(minutes = 5), 0)
df['blackoutMins'] = df['blackoutTime'] / np.timedelta64(1,'m')

这给出了 4 个额外的列

TimeDelta               blackout  blackoutIime           blackoutMins
0 days 00:04:57.310512000   0   0 days 00:00:00.000000000   0.0
0 days 00:05:03.583779000   1   0 days 00:00:03.583779000   0.05972965
0 days 00:15:45.704232000   1   0 days 00:10:45.704232000   10.7617372
0 days 00:04:55.079273000   0   0 days 00:00:00.000000000   0.0
0 days 00:09:10.651137000   1   0 days 00:04:10.651137000   4.17751895
0 days 00:05:05.077309000   1   0 days 00:00:05.077309000   0.08462181666666667
0 days 00:10:59.274865000   1   0 days 00:05:59.274865000   5.9879144166666665
0 days 00:05:26.800226000   1   0 days 00:00:26.800226000   0.44667043333333334
0 days 00:04:58.936685000   0   0 days 00:00:00.000000000   0.0
0 days 00:05:16.684317000   1   0 days 00:00:16.684317000   0.27807195
0 days 00:05:02.304786000   1   0 days 00:00:02.304786000   0.0384131

所以我想要的是可视化 x 轴上的时间停电和 y 轴上的停电，我想要类似的东西

x轴为时间轴，y轴为停电时间。谁能帮忙做一下这个可视化。

Answer 1

你想要 plt.step 反对原来的 timestamp:

df['blackout'] = df.timestamp.diff().gt('5min').astype(int)

plt.step(df.timestamp, df.blackout, c='red')

输出：

根据时间增量值生成图

generate a plot based on time-delta value

python

visualization

time-series

matplotlib