Scribd
Upload
88 views18 pages

Credit Card Customer Analysis

The document presents the results of a credit card customer analysis conducted by Dataking Ltd. It includes an exploration of an dataset of 8,500 customers over a 6-month period. Customers were clustered into 7 groups using k-means clustering. Results show patterns in purchasing behaviors and credit usage across clusters. Recommendations are provided targeted to each cluster, such as increasing credit limits, promoting installment plans, and launching rewards programs to incentivize more purchases.

Uploaded by

Betka Ab
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
88 views18 pages

Credit Card Customer Analysis

The document presents the results of a credit card customer analysis conducted by Dataking Ltd. It includes an exploration of an dataset of 8,500 customers over a 6-month period. Customers were clustered into 7 groups using k-means clustering. Results show patterns in purchasing behaviors and credit usage across clusters. Recommendations are provided targeted to each cluster, such as increasing credit limits, promoting installment plans, and launching rewards programs to incentivize more purchases.

Uploaded by

Betka Ab
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 18

Dataking Ltd.

Eva Polakova, Betka Bohinikova, Michael Flury, Stefan Schultze

Credit card
customer analysis
Table of contents

01
Data exploration
02
Client assumptions
03
Methodology
04
Results & Insights
05
Recommendations
01 - Data exploration

• Dataset received from the


credit card company

​
• 8'500 customers over a

6-month period​
02 - Client assumptions

Amount Number N. and amount Cash advance


of purchases of purchases of installments usage
(cash withdrawals vs. limit)

01 02 03 04
03 - Methodology

Final data processing after having run various analysis to


evaluate different models and select the optimal flow

KNN Imputer Log scaling


Filling missing data points based Scaling data for analysis
on similar customers computation and visualization

K-mean w/ K=7
Identifying and splitting data into
7 distinct customer groups
04 - R&I

Results & Insights


Number of clusters based on metrics

• Distortion Score Elbow

for KMeans Clustering​

​
• K = 5 suggested​
04

Silhouette Plot of KMeans K=5

• K = 5 used as suggested

​
• 8500 samples in 5 centers

04

Silhouette Plot of KMeans K=7 trial

• K increased to 7 centers

​
• 7 well-defined clusters and
increased avg. Silhouette Score
04

Number of clusters based on the t-SNE dimensions reduction

• T-distributed Stochastic

Neighbor Embedding

04

Impact and trends on features between groups

• Establishing features' impact

patterns across clusters​

04

Median values for each cluster group


Mean values for each cluster group
O5 - Recommendations

Customer type 0 Customer type 1

• Uses CC mainly to withdraw cash • Often max out the CC


• Never does purchases • Doesn't use installments
• Uses installments

Recommendation Recommendation
04

• May apply for credit limit increase • Share installments options and advantages

Recommendations

Customer type 2 Customer type 3

• Often max out the CC • Uses CC mainly to withdraw cash


• Purchases frequently • Doesn't use installments
• Uses installments

Recommendation Recommendation
04

• Good customer • Launch points program to increase the


• No recommendation needed number of purchases

Recommendations

Customer type 4 Customer type 5

• Uses CC mainly to withdraw cash • Often max out the CC


• Uses CC to make purchases • Does not make purchases
• Uses installments • Uses installments

Recommendation Recommendation
04

• Good customer • Increase credit limit if eligible


• No recommendation needed

Recommendations

Customer type 6

• Uses CC mainly to withdraw cash


• Does not make purchases
• Does not use installments

Recommendation
04

• Show how installments work


• Advertise loans at low interest rate

Dataking Ltd.
Eva Polakova, Betka Bohinikova, Michael Flury, Stefan Schultze

[email protected]

Thank you

You might also like