通过 Python 中的插值法填充缺失数据

Question

我有一个 pandas 数据框，如下所示：

           Date and Time      Seconds  Pressure (mmHg)  Temperature (C)
0    2021-05-13 13:00:00        0.000          709.719           26.551
1    2021-05-13 14:00:00     3600.001          709.364           25.966
2    2021-05-13 15:00:00     7200.001          708.698           25.331
3    2021-05-13 16:00:00    10800.001          707.689           25.184
4    2021-05-13 17:00:00    14400.001          707.206           25.184

压力和温度数据本应以 15 分钟为间隔，但传感器设置错误，每小时收集一次数据。假设线性插值，如何将数据时间戳扩展到 15 分钟间隔并使用线性插值填充小时之间的缺失数据？我尝试了 here 建议的解决方案，但我的文件很大而且很多。这个解决方案不是很快。

Answer 1

使用DataFrame.resample with Resampler.first for missing values between hours and then DataFrame.interpolate:

df['Date and Time'] = pd.to_datetime(df['Date and Time'])

df = (df.resample('15Min', on='Date and Time')[['Pressure (mmHg)','Temperature (C)']]
       .first()
       .interpolate())
print (df)

                     Pressure (mmHg)  Temperature (C)
Date and Time                                        
2021-05-13 13:00:00        709.71900         26.55100
2021-05-13 13:15:00        709.63025         26.40475
2021-05-13 13:30:00        709.54150         26.25850
2021-05-13 13:45:00        709.45275         26.11225
2021-05-13 14:00:00        709.36400         25.96600
2021-05-13 14:15:00        709.19750         25.80725
2021-05-13 14:30:00        709.03100         25.64850
2021-05-13 14:45:00        708.86450         25.48975
2021-05-13 15:00:00        708.69800         25.33100
2021-05-13 15:15:00        708.44575         25.29425
2021-05-13 15:30:00        708.19350         25.25750
2021-05-13 15:45:00        707.94125         25.22075
2021-05-13 16:00:00        707.68900         25.18400
2021-05-13 16:15:00        707.56825         25.18400
2021-05-13 16:30:00        707.44750         25.18400
2021-05-13 16:45:00        707.32675         25.18400
2021-05-13 17:00:00        707.20600         25.18400

通过 Python 中的插值法填充缺失数据

Filling missing data by interpolation in Python

python

interpolation

missing-data

pandas