Missing time related data

import pandas as pd
import numpy as np
import json

f = open('glucose.json')
dj = json.load(f)
glucose = pd.read_json(dj, orient='records')
glucose.head()

	ID	time	recordtype	glucose
0	2845.0	2019-04-25 00:08	1	109
1	2850.0	2019-04-25 00:50	1
2	2877.0	2019-04-25 07:02	1	123
3	2881.0	2019-04-25 07:34	1	158
4	2886.0	2019-04-25 08:19	1

glucose.head()

	ID	time	recordtype	glucose
0	2845.0	2019-04-25 00:08	1	109
1	2850.0	2019-04-25 00:50	1
2	2877.0	2019-04-25 07:02	1	123
3	2881.0	2019-04-25 07:34	1	158
4	2886.0	2019-04-25 08:19	1

glucose['glucose'] = pd.to_numeric(glucose['glucose'], errors='coerce')
glucose['time'] = pd.to_datetime(glucose['time'], errors='coerce')
glucose['ID'] = glucose['ID'].astype(int)

glucose.dtypes

ID                     int64
time          datetime64[ns]
recordtype             int64
glucose              float64
dtype: object

glucose.head()

	ID	time	recordtype	glucose
0	2845	2019-04-25 00:08:00	1	109.0
1	2850	2019-04-25 00:50:00	1	NaN
2	2877	2019-04-25 07:02:00	1	123.0
3	2881	2019-04-25 07:34:00	1	158.0
4	2886	2019-04-25 08:19:00	1	NaN

glucose = glucose.set_index(['time'])
glucose = glucose.sort_index()
glucose = glucose.reset_index()
glucose.head()

	time	ID	recordtype	glucose
0	2019-04-25 00:08:00	2845	1	109.0
1	2019-04-25 00:14:00	-9223372036854775808	0	NaN
2	2019-04-25 00:29:00	-9223372036854775808	0	NaN
3	2019-04-25 00:44:00	-9223372036854775808	0	NaN
4	2019-04-25 00:50:00	2850	1	NaN

glucose.groupby('recordtype')['glucose'].mean()

recordtype
0           NaN
1    123.403846
Name: glucose, dtype: float64

len(glucose)

136

glucose.glucose.isna().mean()

0.6176470588235294

glucose['interpolated']= glucose['glucose'].interpolate(method='linear')
glucose.head()

	time	ID	recordtype	glucose	interpolated
0	2019-04-25 00:08:00	2845	1	109.0	109.000000
1	2019-04-25 00:14:00	-9223372036854775808	0	NaN	109.466667
2	2019-04-25 00:29:00	-9223372036854775808	0	NaN	109.933333
3	2019-04-25 00:44:00	-9223372036854775808	0	NaN	110.400000
4	2019-04-25 00:50:00	2850	1	NaN	110.866667

full = glucose[glucose['glucose'] > 0]
empty = glucose[glucose['glucose'].isna()]
glucose['roll'] = glucose.interpolated.rolling(10).mean()
glucose.head(10)

	time	ID	recordtype	glucose	interpolated	roll
0	2019-04-25 00:08:00	2845	1	109.0	109.000000	NaN
1	2019-04-25 00:14:00	-9223372036854775808	0	NaN	109.466667	NaN
2	2019-04-25 00:29:00	-9223372036854775808	0	NaN	109.933333	NaN
3	2019-04-25 00:44:00	-9223372036854775808	0	NaN	110.400000	NaN
4	2019-04-25 00:50:00	2850	1	NaN	110.866667	NaN
5	2019-04-25 00:59:00	-9223372036854775808	0	NaN	111.333333	NaN
6	2019-04-25 01:14:00	-9223372036854775808	0	NaN	111.800000	NaN
7	2019-04-25 01:29:00	-9223372036854775808	0	NaN	112.266667	NaN
8	2019-04-25 01:44:00	-9223372036854775808	0	NaN	112.733333	NaN
9	2019-04-25 01:59:00	-9223372036854775808	0	NaN	113.200000	111.1

import matplotlib.pyplot as plt 

plt.figure(figsize=(8, 6))

plt.ylabel('glucose level (mg/dL)')
plt.xlabel('tijd')

plt.plot(glucose['roll'], label = 'rolling mean')
plt.plot(full['glucose'], label = 'gemeten waarde')
plt.plot(empty['interpolated'], 'ro', label = 'geinterpoleerde waarden')
plt.legend()
plt.show()