#!/usr/bin/env python
# coding: utf-8

# # Simple Timeseries Plots in `pandas`
# 
# Using a dummy dataset...

# First, let's create a function to generate a random datetime between two datetimes:

# In[30]:


#https://stackoverflow.com/a/553448/454773
from random import randrange
from datetime import timedelta
from datetime import datetime

def random_datetime(start=datetime.strptime('1/1/2019 1:30 PM','%d/%m/%Y %I:%M %p'),
                end=datetime.strptime('31/1/2019 4:50 AM', '%d/%m/%Y %I:%M %p')):
    """
    This function will return a random datetime between two datetime 
    objects.
    """
    delta = end - start
    int_delta = (delta.days * 24 * 60 * 60) + delta.seconds
    random_second = randrange(int_delta)
    return start + timedelta(seconds=random_second)

random_datetime()


# Now we can createa dummy dataframe with a couple of columns of random data:

# In[63]:


import numpy as np

numsamples = 1000

df=pd.DataFrame({'val1':np.random.rand(numsamples)})
df['datetime']= [random_datetime() for i in range(numsamples)]
#Add some periodicity to the second column of random numbers...
df['val2']=df.apply(lambda x: x['val1']*(1+np.sin(2*np.pi*(x['datetime'].hour-6)/24)), axis=1)
df


# In[64]:


# median value by hour
df.groupby([df['datetime'].dt.hour]).median()


# In[67]:


ax = df.groupby([df['datetime'].dt.hour]).median().reset_index().plot(kind='scatter',
                                                                      x='datetime',y='val1')

df.groupby([df['datetime'].dt.hour]).median().reset_index().plot(kind='scatter',
                                                                 x='datetime',y='val2', color='red', ax=ax)