In [1]:

    

from preamble import *
% matplotlib notebook


    

In [2]:

    

from glob import glob
dfs = []
for file in glob("data/citibike/*.csv"):
    dfs.append(pd.read_csv(file))


    

In [3]:

    

data = pd.concat(dfs)


    

In [4]:

    

data.columns


    

    
Out[4]:





Index(['tripduration', 'starttime', 'stoptime', 'start station id',
       'start station name', 'start station latitude',
       'start station longitude', 'end station id', 'end station name',
       'end station latitude', 'end station longitude', 'bikeid', 'usertype',
       'birth year', 'gender'],
      dtype='object')

In [5]:

    

data.head()


    

    
Out[5]:






  

    

      

      
tripduration
      
starttime
      
stoptime
      
start station id
      
start station name
      
start station latitude
      
start station longitude
      
end station id
      
end station name
      
end station latitude
      
end station longitude
      
bikeid
      
usertype
      
birth year
      
gender
    
  
  

    

      
0
      
326
      
2013-10-01 00:01:08
      
2013-10-01 00:06:34
      
239
      
Willoughby St & Fleet St
      
40.691966
      
-73.981302
      
366
      
Clinton Ave & Myrtle Ave
      
40.693261
      
-73.968896
      
16052
      
Subscriber
      
1982
      
1
    
    

      
1
      
729
      
2013-10-01 00:01:21
      
2013-10-01 00:13:30
      
322
      
Clinton St & Tillary St
      
40.696192
      
-73.991218
      
398
      
Atlantic Ave & Furman St
      
40.691652
      
-73.999979
      
19412
      
Customer
      
\N
      
0
    
    

      
2
      
520
      
2013-10-01 00:01:24
      
2013-10-01 00:10:04
      
174
      
E 25 St & 1 Ave
      
40.738177
      
-73.977387
      
403
      
E 2 St & 2 Ave
      
40.725029
      
-73.990697
      
19645
      
Subscriber
      
1984
      
1
    
    

      
3
      
281
      
2013-10-01 00:01:25
      
2013-10-01 00:06:06
      
430
      
York St & Jay St
      
40.701485
      
-73.986569
      
323
      
Lawrence St & Willoughby St
      
40.692362
      
-73.986317
      
16992
      
Subscriber
      
1985
      
1
    
    

      
4
      
196
      
2013-10-01 00:01:27
      
2013-10-01 00:04:43
      
403
      
E 2 St & 2 Ave
      
40.725029
      
-73.990697
      
401
      
Allen St & Rivington St
      
40.720196
      
-73.989978
      
15690
      
Subscriber
      
1986
      
1
    
  

In [6]:

    

data['one'] = 1
data['starttime'] = pd.to_datetime(data.starttime)
data = data.set_index("starttime")


    

In [7]:

    

data_resampled = data.groupby("start station id").one.resample("3h").sum()


    

In [8]:

    

per_station = data_resampled.unstack(level=0).fillna(0)


    

In [9]:

    

plt.figure()
per_station[301].plot()


    

    















    












    
Out[9]:





<matplotlib.axes._subplots.AxesSubplot at 0x7f3fec194128>

In [10]:

    

from sklearn.gaussian_process import GaussianProcessRegressor


    

In [11]:

    

y = per_station[301].values
X = np.arange(len(y)).reshape(-1, 1)


    

In [12]:

    

gp = GaussianProcessRegressor().fit(X, y)


    

In [13]:

    

plt.figure()
plt.plot(y, label="y")
plt.plot(gp.predict(X), label="preds")
plt.legend()


    

    















    












    
Out[13]:





<matplotlib.legend.Legend at 0x7f3fec060400>

In [14]:

    

gp.kernel_


    

    
Out[14]:





1**2 * RBF(length_scale=1)

In [18]:

    

from sklearn.gaussian_process.kernels import RBF, ExpSineSquared, WhiteKernel
gp = GaussianProcessRegressor(alpha=1, normalize_y=True,
                              kernel = 1.0 * RBF(length_scale_bounds=(2, 500)) + 1.0 * RBF(length_scale_bounds=(50, 1000))
                              + 1.0 * RBF(length_scale=100, length_scale_bounds=(2, 500)) * ExpSineSquared(periodicity=8, periodicity_bounds="fixed")  # + 1.0 * WhiteKernel(noise_level=1)
                              + 1.0 * RBF(length_scale=100, length_scale_bounds=(2, 500)) * ExpSineSquared(periodicity=56, periodicity_bounds="fixed")).fit(X[:1500], y[:1500])


    

In [20]:

    

plt.figure()
plt.plot(y, label="y")
plt.plot(gp.predict(X), label="preds")
plt.legend()


    

    















    












    
Out[20]:





<matplotlib.legend.Legend at 0x7f3feb6c5d68>

In [21]:

    

gp.kernel_


    

    
Out[21]:





1.18**2 * RBF(length_scale=8.34) + 6.15**2 * RBF(length_scale=627) + 5.58**2 * RBF(length_scale=500) * ExpSineSquared(length_scale=0.000298, periodicity=8) + 6.96**2 * RBF(length_scale=37.7) * ExpSineSquared(length_scale=0.0621, periodicity=56)