TimeDelta.ipynb

Kernel: Python 3 (Anaconda)

Bottom plot shows CO2_ppm per min ( with accurate time difference )

Next Steps:

• Converting CO2 per min into Infiltration Rate

• Think of ways to narrow this down. (of course negative slopes), could do by day of week, time span, or values. (probably all three)

import pandas as pd
import numpy as np

df = pd.read_csv('InfiltrationCleanedData.csv', parse_dates = True)

# df.dtypes

df.head(1)

	Unnamed: 0	CO2	Time
0	1	718.0	2016-02-19 13:27:00

df.drop(df.columns[0], axis = 1, inplace = True)

# df.head()

df.dtypes

CO2     float64
Time     object
dtype: object

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

df.dtypes

CO2            float64
Time    datetime64[ns]
dtype: object

df['Time2'] = df.Time.shift(-1)

df.head(3)

	CO2	Time	Time2
0	718.0	2016-02-19 13:27:00	2016-02-19 13:31:00
1	337.0	2016-02-19 13:31:00	2016-02-19 13:36:00
2	332.0	2016-02-19 13:36:00	2016-02-19 13:41:00

df['TimeDelta'] = df.Time2 - df.Time
df.head(3)

	CO2	Time	Time2	TimeDelta
0	718.0	2016-02-19 13:27:00	2016-02-19 13:31:00	00:04:00
1	337.0	2016-02-19 13:31:00	2016-02-19 13:36:00	00:05:00
2	332.0	2016-02-19 13:36:00	2016-02-19 13:41:00	00:05:00

df.dtypes

CO2                  float64
Time          datetime64[ns]
Time2         datetime64[ns]
TimeDelta    timedelta64[ns]
dtype: object

df['TimeDelta'] = df.TimeDelta /  np.timedelta64(1, 'm')

df.dtypes

CO2                 float64
Time         datetime64[ns]
Time2        datetime64[ns]
TimeDelta           float64
dtype: object

df.head()

	CO2	Time	Time2	TimeDelta
0	718.0	2016-02-19 13:27:00	2016-02-19 13:31:00	4.0
1	337.0	2016-02-19 13:31:00	2016-02-19 13:36:00	5.0
2	332.0	2016-02-19 13:36:00	2016-02-19 13:41:00	5.0
3	328.0	2016-02-19 13:41:00	2016-02-19 13:46:00	5.0
4	307.0	2016-02-19 13:46:00	2016-02-19 13:51:00	5.0

df['CO2_2'] = df.CO2.shift(-1)

df.head()

	CO2	Time	Time2	TimeDelta	CO2_2
0	718.0	2016-02-19 13:27:00	2016-02-19 13:31:00	4.0	337.0
1	337.0	2016-02-19 13:31:00	2016-02-19 13:36:00	5.0	332.0
2	332.0	2016-02-19 13:36:00	2016-02-19 13:41:00	5.0	328.0
3	328.0	2016-02-19 13:41:00	2016-02-19 13:46:00	5.0	307.0
4	307.0	2016-02-19 13:46:00	2016-02-19 13:51:00	5.0	296.0

df['CO2_diff'] = df.CO2_2 - df.CO2

#df.head()

df['CO2_per_min'] = df.CO2_diff / df.TimeDelta

df.head(3)

	CO2	Time	Time2	TimeDelta	CO2_2	CO2_diff	CO2_per_min
0	718.0	2016-02-19 13:27:00	2016-02-19 13:31:00	4.0	337.0	-381.0	-95.25
1	337.0	2016-02-19 13:31:00	2016-02-19 13:36:00	5.0	332.0	-5.0	-1.00
2	332.0	2016-02-19 13:36:00	2016-02-19 13:41:00	5.0	328.0	-4.0	-0.80

df.CO2_per_min.plot()

<matplotlib.axes._subplots.AxesSubplot at 0x7f59a863a780>

df.CO2_per_min.describe()

count    1.009840e+05
mean              NaN
std               NaN
min              -inf
25%     -1.200000e+00
50%      0.000000e+00
75%      1.000000e+00
max               inf
Name: CO2_per_min, dtype: float64

# no problems here
# df.TimeDelta.describe()

# Task: FIX THIS
df.CO2_diff.describe()

count    100984.000000
mean         -0.002357
std          14.886084
min        -722.000000
25%          -6.000000
50%           0.000000
75%           5.000000
max         439.000000
Name: CO2_diff, dtype: float64

df.to_csv('Deltas_rough.csv')