In [53]:

    

%matplotlib inline
import pandas as pd
from jupyterworkflow.data import get_fremont_data
data = get_fremont_data()
pivoted = data.pivot_table("Total", index=data.index.time, columns=data.index.date)
pivoted.plot(legend=False, alpha=0.01)


    

    
Out[53]:





<matplotlib.axes._subplots.AxesSubplot at 0x11e909e48>






    

In [54]:

    

pivoted.index


    

    
Out[54]:





Index([00:00:00, 01:00:00, 02:00:00, 03:00:00, 04:00:00, 05:00:00, 06:00:00,
       07:00:00, 08:00:00, 09:00:00, 10:00:00, 11:00:00, 12:00:00, 13:00:00,
       14:00:00, 15:00:00, 16:00:00, 17:00:00, 18:00:00, 19:00:00, 20:00:00,
       21:00:00, 22:00:00, 23:00:00],
      dtype='object')

In [55]:

    

data.index


    

    
Out[55]:





DatetimeIndex(['2012-10-03 00:00:00', '2012-10-03 01:00:00',
               '2012-10-03 02:00:00', '2012-10-03 03:00:00',
               '2012-10-03 04:00:00', '2012-10-03 05:00:00',
               '2012-10-03 06:00:00', '2012-10-03 07:00:00',
               '2012-10-03 08:00:00', '2012-10-03 09:00:00',
               ...
               '2017-07-31 14:00:00', '2017-07-31 15:00:00',
               '2017-07-31 16:00:00', '2017-07-31 17:00:00',
               '2017-07-31 18:00:00', '2017-07-31 19:00:00',
               '2017-07-31 20:00:00', '2017-07-31 21:00:00',
               '2017-07-31 22:00:00', '2017-07-31 23:00:00'],
              dtype='datetime64[ns]', name='Date', length=42312, freq=None)

In [56]:

    

import numpy as np
np.unique(data.index.time)


    

    
Out[56]:





array([datetime.time(0, 0), datetime.time(1, 0), datetime.time(2, 0),
       datetime.time(3, 0), datetime.time(4, 0), datetime.time(5, 0),
       datetime.time(6, 0), datetime.time(7, 0), datetime.time(8, 0),
       datetime.time(9, 0), datetime.time(10, 0), datetime.time(11, 0),
       datetime.time(12, 0), datetime.time(13, 0), datetime.time(14, 0),
       datetime.time(15, 0), datetime.time(16, 0), datetime.time(17, 0),
       datetime.time(18, 0), datetime.time(19, 0), datetime.time(20, 0),
       datetime.time(21, 0), datetime.time(22, 0), datetime.time(23, 0)], dtype=object)

In [57]:

    

!head -24 Fremont.csv


    

    




Date,Fremont Bridge West Sidewalk,Fremont Bridge East Sidewalk
10/03/2012 12:00:00 AM,4,9
10/03/2012 01:00:00 AM,4,6
10/03/2012 02:00:00 AM,1,1
10/03/2012 03:00:00 AM,2,3
10/03/2012 04:00:00 AM,6,1
10/03/2012 05:00:00 AM,21,10
10/03/2012 06:00:00 AM,105,50
10/03/2012 07:00:00 AM,257,95
10/03/2012 08:00:00 AM,291,146
10/03/2012 09:00:00 AM,172,104
10/03/2012 10:00:00 AM,72,46
10/03/2012 11:00:00 AM,10,32
10/03/2012 12:00:00 PM,35,41
10/03/2012 01:00:00 PM,42,48
10/03/2012 02:00:00 PM,77,51
10/03/2012 03:00:00 PM,72,92
10/03/2012 04:00:00 PM,133,182
10/03/2012 05:00:00 PM,192,391
10/03/2012 06:00:00 PM,122,258
10/03/2012 07:00:00 PM,59,69
10/03/2012 08:00:00 PM,29,51
10/03/2012 09:00:00 PM,25,38
10/03/2012 10:00:00 PM,24,25

In [58]:

    

x = pivoted.fillna(0).T.values
x.shape


    

    
Out[58]:





(1763, 24)

In [59]:

    

from sklearn.decomposition import PCA
x2 = PCA(2, svd_solver="full").fit_transform(x)


    

In [60]:

    

x2.shape


    

    
Out[60]:





(1763, 2)

In [61]:

    

import matplotlib.pyplot as plt
plt.scatter(x2[:, 0], x2[:, 1])


    

    
Out[61]:





<matplotlib.collections.PathCollection at 0x120cd0320>






    

In [62]:

    

from sklearn.mixture import GaussianMixture
gmm = GaussianMixture(2)
gmm.fit(x)
labels = gmm.predict(x)
labels


    

    
Out[62]:





array([1, 1, 1, ..., 0, 0, 1])

In [63]:

    

plt.scatter(x2[:, 0], x2[:, 1], c=labels)
plt.colorbar()


    

    
Out[63]:





<matplotlib.colorbar.Colorbar at 0x120e39e48>






    

In [64]:

    

plt.scatter(x2[:, 0], x2[:, 1], c=labels, cmap="rainbow")
plt.colorbar()


    

    
Out[64]:





<matplotlib.colorbar.Colorbar at 0x120cba358>






    

In [65]:

    

pivoted.T[labels == 0].T.plot(legend=False, alpha=0.01)


    

    
Out[65]:





<matplotlib.axes._subplots.AxesSubplot at 0x120d372e8>






    

In [66]:

    

pivoted.T[labels == 1].T.plot(legend=False, alpha=0.01)


    

    
Out[66]:





<matplotlib.axes._subplots.AxesSubplot at 0x121b9db00>






    

In [67]:

    

pd.DatetimeIndex(pivoted.columns).dayofweek


    

    
Out[67]:





Int64Index([2, 3, 4, 5, 6, 0, 1, 2, 3, 4,
            ...
            5, 6, 0, 1, 2, 3, 4, 5, 6, 0],
           dtype='int64', length=1763)

In [68]:

    

dayofweek = pd.DatetimeIndex(pivoted.columns).dayofweek


    

In [70]:

    

plt.scatter(x2[:, 0], x2[:, 1], c=dayofweek, cmap="rainbow")
plt.colorbar()


    

    
Out[70]:





<matplotlib.colorbar.Colorbar at 0x12340ffd0>






    

In [81]:

    

dates = pd.DatetimeIndex(pivoted.columns)
dates[(labels ==0) & (dayofweek < 1)]


    

    
Out[81]:





DatetimeIndex(['2012-12-24', '2013-05-27', '2013-09-02', '2014-05-26',
               '2014-09-01', '2015-05-25', '2015-09-07', '2016-05-30',
               '2016-07-04', '2016-09-05', '2016-12-26', '2017-01-02',
               '2017-02-06', '2017-05-29'],
              dtype='datetime64[ns]', freq=None)

In [ ]: