Vanya Sahu

Vanya Sahu

Data Scientist | B.Tech - CSE

Basics : Linear Regression

3 minute read

Objective: Given the customer data of a company, whether it should focus on their website or mobile app experience to increase sales.

Source: Udemy | Python for Data Science and Machine Learning Bootcamp
Data used in the below analysis: link

#Import all the libraries for data analysis
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
#used to view plots within jupyter notebook
%matplotlib inline
sns.set_style("whitegrid") #setting view for plots, optional

customers = pd.read_csv('Ecommerce Customers') #import dataset

customers.head(2) #view dataset
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 
#view some informormation of the dataset
customers.describe()
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 
customers.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 500 entries, 0 to 499
Data columns (total 8 columns):
 #   Column                Non-Null Count  Dtype  
---  ------                --------------  -----  
 0   Email                 500 non-null    object
 1   Address               500 non-null    object
 2   Avatar                500 non-null    object
 3   Avg. Session Length   500 non-null    float64
 4   Time on App           500 non-null    float64
 5   Time on Website       500 non-null    float64
 6   Length of Membership  500 non-null    float64
 7   Yearly Amount Spent   500 non-null    float64
dtypes: float64(5), object(3)
memory usage: 31.4+ KB

#View numerics data in the dataset and the relationship with other columns
sns.pairplot(data=customers)

Pair Plot

It seems Lenth of Membership is most correlated with Yearly Amount Spent

#Playing aorund with the data to check for more correlations
sns.jointplot(x='Time on Website',y='Yearly Amount Spent',data=customers)

Pair Plot

sns.jointplot(x='Time on App',y='Yearly Amount Spent',data=customers)

Pair Plot

sns.jointplot(x='Time on App',y='Length of Membership',data=customers,kind='hex')

Pair Plot

#Can see the correlation clearly here
sns.lmplot(x='Length of Membership',y='Yearly Amount Spent', data = customers)

Pair Plot

Now we start training the model to predict Yearly Amount spent based on given information

#view column names in the dataset
customers.columns

Index(['Email', 'Address', 'Avatar', 'Avg. Session Length', 'Time on App',
       'Time on Website', 'Length of Membership', 'Yearly Amount Spent'],
      dtype='object')

#divide data into X and y DataFrames
X = customers[['Avg. Session Length', 'Time on App',
       'Time on Website', 'Length of Membership']]
y = customers['Yearly Amount Spent']
#Split the X and y into Training and Test
from sklearn.model_selection import train_test_split
X_train,X_test, y_train,y_test =train_test_split(X, y, test_size=0.3, random_state=101)

#import Linear Regression model from Sci-kit Learn
from sklearn.linear_model import LinearRegression

lm = LinearRegression()
lm.fit(X_train,y_train)

LinearRegression()

#view coefficients, or amount of correlation among Yearly Amount spent and other data provided
lm.coef_

array([25.98154972, 38.59015875,  0.19040528, 61.27909654])

#predict the values on the test data
predictions = lm.predict(X_test)

#plot the result alongwith the actual data
plt.scatter(y_test,predictions)

Pair Plot

Model rendered seems like a good fit

#import metrics to evaluate model performance
from sklearn import metrics

print('MAE:', metrics.mean_absolute_error(y_test, predictions))
print('MSE:', metrics.mean_squared_error(y_test, predictions))
print('RMSE:', np.sqrt(metrics.mean_squared_error(y_test, predictions)))

MAE: 7.228148653430839
MSE: 79.81305165097469
RMSE: 8.933815066978648

#plot the residual amount to check model performance
sns.distplot(y_test-predictions,bins=30)

Pair Plot

Check the correlation with coefficients seen before

Coeff = pd.DataFrame(lm.coef_,X.columns,columns=['Coefficients'])
Coeff
Coefficients
Avg. Session Length 25.981550
Time on App 38.590159
Time on Website 0.190405
Length of Membership 61.279097

The above data signify that 1 change in Avg. Session Length will lead to 25.98 change in Yearly Amount Spent.
Similarly, for Time on App, Time on Website and Length of Membership
As predicted before, the most impact if from Length of Membership

Result: The most correlated field is Length of Membership.

However, we need to determine the focus of the company on ehancing the Website / App for users. This could be done depending on time and resources and the approach the company would like to take. Here are three options:

1. Enhance Website as it doesn’t help in the sales right now.

2. Enhance App for even better performance.

3. Look for other metrics as the most sales are coming in due to length of membership and neither from Website nor App.

You May Also Enjoy