Linear Regression Project¶

Ecommerce company is trying to decide whether to focus their efforts on their mobile app experience or their website. Let's figure it out!

In [1]:
# Data and Setup.
In [2]:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
In [3]:
customers = pd.read_csv('Ecommerce Customers')
In [4]:
customers.head()
Out[4]:
Email Address Avatar Avg. Session Length Time on App Time on Website Length of Membership Yearly Amount Spent
0 mstephenson@fernandez.com 835 Frank Tunnel\nWrightmouth, MI 82180-9605 Violet 34.497268 12.655651 39.577668 4.082621 587.951054
1 hduke@hotmail.com 4547 Archer Common\nDiazchester, CA 06566-8576 DarkGreen 31.926272 11.109461 37.268959 2.664034 392.204933
2 pallen@yahoo.com 24645 Valerie Unions Suite 582\nCobbborough, D... Bisque 33.000915 11.330278 37.110597 4.104543 487.547505
3 riverarebecca@gmail.com 1414 David Throughway\nPort Jason, OH 22070-1220 SaddleBrown 34.305557 13.717514 36.721283 3.120179 581.852344
4 mstephens@davidson-herman.com 14023 Rodriguez Passage\nPort Jacobville, PR 3... MediumAquaMarine 33.330673 12.795189 37.536653 4.446308 599.406092
In [5]:
customers.describe()
Out[5]:
Avg. Session Length Time on App Time on Website Length of Membership Yearly Amount Spent
count 500.000000 500.000000 500.000000 500.000000 500.000000
mean 33.053194 12.052488 37.060445 3.533462 499.314038
std 0.992563 0.994216 1.010489 0.999278 79.314782
min 29.532429 8.508152 33.913847 0.269901 256.670582
25% 32.341822 11.388153 36.349257 2.930450 445.038277
50% 33.082008 11.983231 37.069367 3.533975 498.887875
75% 33.711985 12.753850 37.716432 4.126502 549.313828
max 36.139662 15.126994 40.005182 6.922689 765.518462
In [6]:
customers.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 500 entries, 0 to 499
Data columns (total 8 columns):
Email                   500 non-null object
Address                 500 non-null object
Avatar                  500 non-null object
Avg. Session Length     500 non-null float64
Time on App             500 non-null float64
Time on Website         500 non-null float64
Length of Membership    500 non-null float64
Yearly Amount Spent     500 non-null float64
dtypes: float64(5), object(3)
memory usage: 31.3+ KB
In [7]:
# Exploring data.
In [8]:
# Jointplots - Time on Website to Yearly Amount Spent and Time on App to Yearly Amount Spent.
In [10]:
sns.set_palette('GnBu_d')
sns.set_style('whitegrid')
sns.jointplot(x='Time on Website', y='Yearly Amount Spent',data=customers)
Out[10]:
<seaborn.axisgrid.JointGrid at 0x28a2a965ef0>
In [11]:
sns.jointplot(x='Time on App', y='Yearly Amount Spent',data=customers)
Out[11]:
<seaborn.axisgrid.JointGrid at 0x28a2aa2dc88>

More time on site, more money spent.

In [12]:
# Hex jointplot - Time on App and Length of Membership.
In [13]:
sns.jointplot(x='Time on App',y='Length of Membership',kind='hex',data=customers)
Out[13]:
<seaborn.axisgrid.JointGrid at 0x28a2abe1a58>
In [14]:
# Exploring these types of relationships across the entire data set.
In [15]:
sns.set_palette('dark')
sns.pairplot(customers, height=1.8)
Out[15]:
<seaborn.axisgrid.PairGrid at 0x28a2ad3e5c0>

Length of Membership is the most correlated feature with Yearly Amount Spent.

In [16]:
# Linear model plot - Yearly Amount Spent vs. Length of Membership.
In [17]:
sns.lmplot(x='Length of Membership',y='Yearly Amount Spent',data=customers, height=8, palette='deep')
Out[17]:
<seaborn.axisgrid.FacetGrid at 0x28a2b673f28>
In [18]:
# Training and Testing Data.
In [19]:
customers.columns
Out[19]:
Index(['Email', 'Address', 'Avatar', 'Avg. Session Length', 'Time on App',
       'Time on Website', 'Length of Membership', 'Yearly Amount Spent'],
      dtype='object')
In [20]:
y = customers['Yearly Amount Spent']
In [21]:
X = customers[['Avg. Session Length', 'Time on App','Time on Website', 'Length of Membership']]
In [22]:
from sklearn.model_selection import train_test_split
In [23]:
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=101)
In [24]:
# Training the Model.
In [25]:
from sklearn.linear_model import LinearRegression
In [26]:
lm = LinearRegression()
In [27]:
lm.fit(X_train, y_train)
Out[27]:
LinearRegression(copy_X=True, fit_intercept=True, n_jobs=None,
         normalize=False)
In [28]:
# What are the coefficients of the model?
In [29]:
print('Coefficients: ', lm.coef_)

Coefficients:  [25.98154972 38.59015875  0.19040528 61.27909654]
In [30]:
# Predicting Test Data.
In [31]:
predictions = lm.predict(X_test)
In [32]:
plt.figure(figsize=(12, 8))
plt.scatter(y_test,predictions,edgecolors='black',c='g')
plt.xlabel('Y Test')
plt.ylabel('Predicted Y')
Out[32]:
Text(0, 0.5, 'Predicted Y')
In [33]:
# Evaluating the Model.
In [34]:
from sklearn import metrics

print('Mean Absolute Error:', metrics.mean_absolute_error(y_test, predictions))
print('Mean Squared Error:', metrics.mean_squared_error(y_test, predictions))
print('Root Mean Squared Error:', np.sqrt(metrics.mean_squared_error(y_test, predictions)))

Mean Absolute Error: 7.228148653430853
Mean Squared Error: 79.81305165097487
Root Mean Squared Error: 8.933815066978656
In [35]:
print('Explained Variance Score:', metrics.explained_variance_score(y_test, predictions))

Explained Variance Score: 0.9890771231889606
In [36]:
 # Residuals
In [38]:
plt.figure(figsize=(11, 5))
sns.distplot((y_test-predictions),bins=50);
In [39]:
# Conclusion
In [40]:
coeffecients = pd.DataFrame(lm.coef_,X.columns)
coeffecients.columns = ['Coeffecient']
coeffecients
Out[40]:
Coeffecient
Avg. Session Length 25.981550
Time on App 38.590159
Time on Website 0.190405
Length of Membership 61.279097
  • Holding all other features fixed, a 1 unit increase in Avg. Session Length is associated with an increase of £25.98 total spent.
  • Holding all other features fixed, a 1 unit increase in Time on App is associated with an increase of £38.59 total spent.
  • Holding all other features fixed, a 1 unit increase in Time on Website is associated with an increase of £0.19 total spent.
  • Holding all other features fixed, a 1 unit increase in Length of Membership is associated with an increase of £61.27 total spent.
In [41]:
# Does company should focus more on their mobile app or website development?

There are two ways to think about this: Develop the Website to catch up to the performance of the mobile app, or develop the app more since that is what is working better.

In [42]:
#The End