Week4_portfolio
docx
keyboard_arrow_up
School
Johns Hopkins University *
*We aren’t endorsed by this school
Course
BU.510.601
Subject
Information Systems
Date
Dec 6, 2023
Type
docx
Pages
3
Uploaded by CoachField12328
How to identify potential membership customers by payment preference in the
supermarket?
Supermarkets like Costco and Sam require customers to join their membership in order to shop inside, but
not every customer prefers joining the membership due to various reasons. As a group of finance students who are
interested in opening a supermarket near Carey Business School, we are interested in identifying the customers’
preference of joining the supermarket membership in order to retain customers. For the first week, we would like
to explore if the payment method is related to the customers’ intention to join the membership, for example, people
using credit card might be more willing to join the membership since they might have greater desire for
consumption. We will be utilizing
Supermarket Sales Data
from Kaggle to answer the following question:
1.
What percentage of customers are members of the supermarket?
2.
What percentage of customers are members of the supermarket and use Ewallet(or credit card) for payment?
3.
Does the probability that customers using Ewallet (or credit card) for payment depending on the type of
customers?
4.
What is the probability that a customer using Ewallet (or credit card) is a member of the supermarket?
Information About Data and Data Preparation
The original data is provided by ARUN JANGIR agreeing on Open Database License 1.0 which provides a
comprehensive overview of supermarket sales, offering insights into consumer purchasing behaviors, product
trends, and retail dynamics. We could explore sales volumes, seasonal patterns, and promotional impacts for
informed decision-making in the competitive retail sector by utilizing this data form. This data form has 12
variables and 1001
rows. To answer the question above, we only need customers’ membership and their payment method information.
This is a well-formed, cleaned data form so we do not need to clean and sort it out, we just removed features that
we don’t need and then we are good to go. You can find the filtered form on our
Github Repository
.
Figure 1 shows the frequency of intervals of 1,000 customers' expenditure in the supermarket. The
horizontal axis is the total goods expenditure. The vertical axis represents the frequency of occurrence. The
histogram may follow a Poisson distribution with a mean of 322.97 dollars.
Figure 1 Customer Expenditure Histogram
Analysis and Results
To answer our 4 research questions, we utilized Excel Function “COUNTIF (range, criteria)” and
“COUNTIFS (range1, criteria1, range2, criteria2)” to count the number of members and those who are using
Ewallet or Credit card for payment. The results are shown in Table 1.
Table 1:Relative Frequency and Conditional Probability Results
The probability that a customer would purchase a membership is 501/1000, which indicates that about half of
the customers are likely to be a member of the supermarket. Among all the 1000 customers, 345 of them are using
Ewallet to pay and only 311 of them are using Credit Card. However, the probability that a customer would use
Ewallet given he/she is a member is 161/501 and the probability that a customer would use credit card given
he/she is a member is 172/501, which indicates that those customers who are members of the supermarket are
more likely to use credit card. In fact, if we are trying to identify customers who are more likely to join our
membership by observing their payment method, conditional probability P(Member|Ewallet) and P(Member|
Credit) is what we are looking for. Applying Bayes’ Theorem to the calculation (as it is calculated above), we have
P(Member|Ewallet) = 161/345 and P(Member|Credit) = 172/311. Comparing the two probabilities, we might draw
to the conclusion that people using credit card to pay for their shopping are more likely to purchase a membership
in the supermarket. Therefore, possible strategies here could be to try to encourage customers to pay by credit card
or try to attract more credit card users to shop in our supermarket, such as giving a discount to credit card
1.
P(Member)=501/1000
2.
P(Ewallet∩Member)=161/1000
P (Credit∩Member) = 172/1000
3.
P(Member)=501/1000 P(Ewallet)=345/1000 P(Credit)=311/1000
P(Ewallet|Member) = 161/501
P(Ewallet|Normal) = 184/499
P(Credit|Member) = 172/501
P(Credit|Normal) = 139/499
4.
P(Member|Ewallet) = P (Member, Ewallet)/P(Ewallet) = 161/345
P(Member|Credit) = P (Member, Credit)/P(Credit) = 172/311
P(Member|Credit) > P(Member|Ewallet)
payments or establish extra points system for credit cards.
Week 4 Progress:
The company hired a marketing research group from Carey Business School, who randomly selected 100
customers from the 1000-person population data and calculated their average expenditure equals to 316.70, under a
95% confidence level, should the company doubt the result of the research group? The manager said he selected a
100-people sample himself and the average customer expenditure is more than 406, is his data reliable?
The overall Customers’ expenditure follows Poisson distribution with a mean of 322.97: u
x
=322.97
σ
x
= 322.97
The random selected sample consists of n=100 people,
= 316.70
x̄
Confidence Interval = (
-Z
x̄
α/2
*σ
x
/sqrt(n), +Z
x̄
α/2
*σ
x
/sqrt(n)) = (316.70-1.96*322.97/10, 316.70+1.96*322.97/10) =
(253.20, 380.20)
Therefore, the company should not doubt the result of the research group under 95% confidence level.
Calculate the manager’s data’s confidence level
P(Z= (406- u
x
)/(σ
x
/sqrt(n)) = P(Z=2.58)=0.01
The manager’s data is unreliable.
(Portfolio updates on GitHub every week
Link
)
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
- Access to all documents
- Unlimited textbook solutions
- 24/7 expert homework help