In [2]:

    

import numpy as np
import matplotlib.pyplot as plt
from matplotlib import rcParams
import datetime as dt
from sklearn import datasets, linear_model
%matplotlib inline


    

In [3]:

    

data = np.genfromtxt('recs2009_public.csv',delimiter=',',skip_header=1)


    

In [4]:

    

totalSF=data[:,827]


    

In [5]:

    

totalKWH=data[:,839]


    

In [6]:

    

income=data[:,785]


    

In [7]:

    

age=data[:,760]


    

In [8]:

    

members=data[:,759]


    

In [9]:

    

education=data[:,758]


    

In [10]:

    

plt.figure(figsize=(15,10))
plt.plot(totalSF,totalKWH,'ro')
plt.ylabel('Residential Electric Consumption [kWH]')


    

    
Out[10]:





<matplotlib.text.Text at 0x7f16a54bc250>






    

In [11]:

    

SFtrain=totalSF[:-6040]
KWHtrain=totalKWH[:-6040]


    

In [12]:

    

print SFtrain


    

    




[ 5075.  3136.   528. ...,   644.   834.   952.]

In [13]:

    

SFtest=totalSF[-6040:]
KWHtest=totalKWH[-6040:]


    

In [14]:

    

print SFtest


    

    




[ 3640.  2025.  3068. ...,  4581.  1728.  4920.]

In [15]:

    

plt.figure(figsize=(15,10))
plt.plot(SFtrain,KWHtrain,'ro')
plt.ylabel('Residential Electric Consumption Train Data [kWH]')
plt.xlabel('Total Residential Area [SF]')


    

    
Out[15]:





<matplotlib.text.Text at 0x7f16ad163b50>






    

In [16]:

    

plt.figure(figsize=(15,10))
plt.plot(SFtest,KWHtest,'ro')
plt.ylabel('Residential Electric Consumption [kWH]')
plt.xlabel('Total Residential Area [SF]')


    

    
Out[16]:





<matplotlib.text.Text at 0x7f16ace7dc50>






    

In [17]:

    

from scipy import stats


    

In [18]:

    

slope, intercept, r_value, p_value, std_err = stats.linregress(SFtrain,KWHtrain)


    

In [19]:

    

slope


    

    
Out[19]:





2.0735207102480704

In [20]:

    

intercept


    

    
Out[20]:





6817.2690298966618

In [21]:

    

KWHpredict=slope*SFtest+intercept


    

In [22]:

    

plt.figure(figsize=(15,10))
plt.plot(SFtest,KWHtest,'.b')
plt.plot(SFtest,KWHpredict,'.r')
plt.legend(['true data','predicted data'])
plt.title('Prediction on testing data')


    

    
Out[22]:





<matplotlib.text.Text at 0x7f16acd4f450>






    

In [46]:

    

plt.scatter(data[:,34], data[:,461])


    

    
Out[46]:





<matplotlib.collections.PathCollection at 0x7f16a863bc50>






    




/opt/anaconda/envs/np18py27-1.9/lib/python2.7/site-packages/matplotlib/collections.py:590: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison
  if self._edgecolors == str('face'):






    

In [29]:

    

full=data[np.where(data[:,34]==data[:,461])]


    

In [31]:

    

SFfull=full[:,827]


    

In [32]:

    

SFfull


    

    
Out[32]:





array([ 5075.,  3136.,   528., ...,  4581.,  1728.,  4920.])

In [35]:

    

KWHfull=full[:,839]


    

In [36]:

    

len(SFfull)


    

    
Out[36]:





9472

In [38]:

    

SFfulltrain=SFfull[:-4718]
SFfulltest=SFfull[-4718:]


    

In [39]:

    

KWHfulltrain=KWHfull[:-4718]
KWHfulltest=KWHfull[-4718:]


    

In [40]:

    

plt.figure(figsize=(15,10))
plt.plot(SFfulltrain,KWHfulltrain,'ro')
plt.ylabel('Residential Electric Consumption Full[kWH]')
plt.xlabel('Total Residential Area [SF]')


    

    
Out[40]:





<matplotlib.text.Text at 0x7f16acbfb190>






    

In [41]:

    

slopef, interceptf, r_valuef, p_valuef, std_errf = stats.linregress(SFfulltrain,KWHfulltrain)


    

In [42]:

    

slopef


    

    
Out[42]:





2.0986028355898076

In [43]:

    

interceptf


    

    
Out[43]:





6939.3870321567492

In [44]:

    

KWHfullpredict=slopef*SFfulltest+interceptf


    

In [45]:

    

plt.figure(figsize=(15,10))
plt.plot(SFfulltest,KWHfulltest,'.b')
plt.plot(SFfulltest,KWHfullpredict,'.r')
plt.legend(['true data','predicted data'])
plt.title('Prediction on testing data')


    

    
Out[45]:





<matplotlib.text.Text at 0x7f16a8724890>






    

In [ ]: