data visualisation

from __future__ import division

import pandas as pd
import matplotlib.pyplot as plt
from datetime import timedelta

%matplotlib inline

SLEEP_INTERVAL = 5 # minutes
# https://www.timeanddate.com/sun/singapore/singapore?month=4&year=2018
SUNRISE_TIME, SUNSET_TIME = '07:04', '19:12'

# https://en.wikipedia.org/wiki/Heart_rate
HR_LOW_LVL = 50
HR_REST_LVL = 100

xl = pd.ExcelFile('./patient_1.xlsx')
xl.sheet_names

data = xl.parse('data')
data['timestamp'] = pd.to_datetime(data.timestamp)
min_datetime = pd.to_datetime('{} 12:00:00'.format(data.timestamp.min().date()))
data['day'] = data.timestamp.apply(lambda x: (x - min_datetime).days)
data = data.sort_values(['day', 'timestamp'])
data.head()


columns = ['Activity', 'hr', 'rr', 'temp', 'sleep', 'timestamp', 'Alarm', 'AF', 'sym', 'day']
df = data[data.day == 1]


def get_sleep_duration(daily_data):
    daily_data['sleep_instance'] = (daily_data.sleep.diff().fillna(0) != 0).cumsum()
    sleep_df = daily_data[lambda x: x.sleep == 1].groupby('sleep_instance')['timestamp'].agg({'start': min, 'end': max})
    sleep_df['timegap'] = (sleep_df['start'] - sleep_df['end'].shift(1)).fillna(timedelta(0)).astype('timedelta64[m]').astype(int)
    sleep_df['sleep_instance'] = (sleep_df['timegap'] > SLEEP_INTERVAL).cumsum()
    durations = []
    for k, v in sleep_df.groupby('sleep_instance'):
        durations.append((v['start'].min(), v['end'].max()))
    return pd.DataFrame(durations, columns=['start', 'end'])

def get_hr_lvl(hr):
    if hr < HR_LOW_LVL:
        return 1 # low
    if (hr >= HR_LOW_LVL) and (hr <= HR_REST_LVL):
        return 2 # normal
    if (hr > HR_REST_LVL):
        return 3 # high
    return 
# def get_hr_duration(daily_data)
daily_data = df.copy()
daily_data['hr_lvl'] = df.hr.apply(get_hr_lvl)
daily_data['hr_instance'] = (daily_data['hr_lvl'].diff().fillna(0) != 0).cumsum()

durations = []
for k , g in daily_data.groupby('hr_instance'):
    start, end = g.timestamp.min(), g.timestamp.max()
    hr_lvl = g['hr_lvl'].mode().iloc[0]
    durations.append((start, end, hr_lvl))
pd.DataFrame(durations, columns=['start', 'end', 'hr_lvl'])


get_sleep_duration(df)