Scribd
Upload
25 views

Lecture 03 - Student Version

This document discusses using regression analysis to evaluate the impact of advertisement spending on sales. It provides an example using ice cream sales data to illustrate how to perform linear regression analysis. Key steps include drawing a scatter plot of the variables, calculating the correlation coefficient, fitting a least squares regression line that minimizes the residuals between predicted and actual values, and using calculus to determine the slope and intercept coefficients.

Uploaded by

yashdugar.fun
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
25 views

Lecture 03 - Student Version

This document discusses using regression analysis to evaluate the impact of advertisement spending on sales. It provides an example using ice cream sales data to illustrate how to perform linear regression analysis. Key steps include drawing a scatter plot of the variables, calculating the correlation coefficient, fitting a least squares regression line that minimizes the residuals between predicted and actual values, and using calculus to determine the slope and intercept coefficients.

Uploaded by

yashdugar.fun
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 15

Lecture 03: Regression

Analysis
Going back to the
marketing meeting..
• Remember working at the Marketing and Analytics division of
Nykaa?
• You propose running more advertisements on the website, this
time featuring a well-noted celebrity to improve Nykaa’s sales.
• & Congratulations, you successfully complete it!
• One fine day, suddenly the management calls you in for a
quarterly meeting, and says:

“Your advertisement costed us Rs. 75 lakhs. Do


you think it worked? Advertisements and
promotions involve a major share of a firm’s
costs. Didn’t you know that? ”
• The task of answering such questions may seem daunting at first.

• What do you think is the kind of data you need to handle such questions?

• At this point, you may only have a vague idea of the kind of data you would need to collect.

• By the end of this course, you should know how to use statistical methods to formally evaluate impact of
advertisement campaigns.
Hands-On Data Exercise: Part 1
• Download the dataset showing Ad data and Nykaa
Sales

• Step 1: Using knowledge from previous lectures!


• As more we spend on Ads, more people get to know about
products/campaigns/sales => sales increase.

• May not be.

• However, using regression analysis, you can predict the


number of orders from the ad spends too.
Goal of Regression
• Obtain the regression equation

• In the form of y = mx + c

• When you input a value for ‘x’, you


can predict how many ‘y’ were sold
Model Specification
sales = β0 + β1 advertisement expenses + μ

Dependent Independent
Error term
variable variable

Regression
Coefficient

Also knows as parameters

describe the directions and


strengths of the relationship
between Ads and Sales

Slope of the line


So, how do you get the regression equation?

Hands-On Data Exercise: Part 2


Ice-Cream Data Q1. Draw Scatter Plot
Q2. Compute Correlation Coefficient

Scatter Plot
• Value of correlation Coefficient = 0.9069, which is
100

90
high
80
Ice Tea Orders

70
• Variables are highly positively correlated
60

50

40
20 22 24 26 28 30 32 34 36

Temp (Celcius)
• Draw a straight line following the pattern in the data
Scatter Plot
100 • Select the graph, click on + sign, select “trendline”
90
option
80
Ice Tea Orders

70 • Little arrows are the distances from the line which


60
represents the estimated values and the dot which are
50

40
20
22 2224 24 26 26 28 28 30 30 32 32 3434 36 the actual values.
Temp (Celcius)
• This distance is the residual.

This is known as linear least


• The goal is to find the line that best minimizes all the
squares regression.
residuals.
Scatter Plot
100 • We square the residuals to find the sum of squares,
90

80
which we use to find the regression equation.
Ice Tea Orders

70

60 • Yi = α + βXi + εi
50

40
22 24 26 28 30 32 34 36 • For a value Xi of the independent variable, the
Temp (Celcius)
estimated value of the dependent variable Y i is given

by Ŷi = α + βXi
This is known as linear least
squares regression. • The difference between actual value Yi and the

estimated value Ŷi is error term/residual

• ei = Yi - Ŷi
• We square the residuals to find the sum of squares, which we use to find the regression equation.

• Using calculus, unknowns ‘a’ and ‘b’ are obtained by equating first derivative of the sum of square of errors to zero.

Raw materials
of step 2
• Using calculus, unknowns ‘a’ and ‘b’ are obtained by equating first derivative of the sum of square of errors to zero.

• Once slope is determined, intercept can be found using the equation


Numerical Calculation: Ice-Cream Data

You might also like