Objective Questions:

1. What is the distribution of account balances across different

regions?
select avg(b.Balance) as Avgbalance,g.GeographyLocation from bank_churn b
join customerinfo c
on b.CustomerId=c.CustomerId
join geography g
on c.GeographyID=g.GeographyID
group by g.GeographyLocation
order by Avgbalance desc;

Joining the three tables, customerinfo, Bankchurn, and Geography,

yields the desired output. The Bankchurn table facilitates obtaining the
average balance through the AVG function, while the Geography table is linked
to the customerinfo table via the GeographyID using the JOIN function. The
GROUP BY function is employed to group the location-based data, and the
ORDER BY function arranges the output in descending order of the average
balance. From the resulting output, it is observed that Germany has the
highest average customer account balance, amounting to
119730.11613391782.
2. Identify the top 5 customers with the highest Estimated Salary in
the last quarter of the year. (SQL)
SELECT
CustomerId,Surname,Age,GenderID,EstimatedSalary,GeographyID,BankDOJ
FROM customerinfo
WHERE MONTH(BankDOJ) IN (10, 11, 12)
ORDER BY EstimatedSalary DESC
LIMIT 5;
In order to retrieve the highest estimated salary in the last quarter
(October, November, and December), the WHERE clause filters data based on
the months. Utilizing the MONTH function, the month is extracted, and the 'IN'
operator specifies the months. Subsequently, the estimated salaries are sorted
in descending order using the ORDER BY clause. Finally, limiting the output to
the top 5 highest estimated salaries is achieved through the LIMIT clause.

Output:

Surname EstimaedSalary BankDOJ

Dyer 199970.74 2016-11-29
Oluchukwu 199841.32 2019-12-25
Mai 199805.63 2018-11-12
Palerma 199753.97 2019-10-01
Dimauro 199638.56 2018-11-16
3. Calculate the average number of products used by customers
who have a credit card. (SQL)
SELECT round(avg(NumofProducts),2) as Avg_num_of_products from
bank_churn
where HasCrCard='1';
The average number of products can be calculated using the AVERAGE
function, with a WHERE condition applied to filter customers who have a credit
card.

Output:

Avg_num_of_products
1.53

4. Determine the churn rate by gender for the most recent year in
the dataset.
In the analysis conducted using a table chart, the churn rate by gender
for the most recent year available in the dataset, 2019, was determined.
Notably, the churn rate for females (25.05%) exceeded that of males (15.37%),
despite the total count of exited customers being 658. It's noteworthy that the
total number of male customers is higher than females, yet the count of
exited customers is lower compared to females.

Output:

5. Compare the average credit score of customers who have exited

 Avg_num_of_products
1.53

and those who remain. (SQL)

SELECT ec.ExitCategory,
AVG(bc.CreditScore) AS AvgCreditScore
FROM Exitcustomer ec
JOIN bank_churn bc
ON ec.ExitID = bc.Exited
GROUP BY ec.ExitCategory;
This SQL query calculates the average credit score of customers grouped by
their exit category. It selects the exit category from the Exitcustomer table and
the average credit score from the CreditScore column of the bank_churn table,
aliasing it as AvgCreditScore. The query joins the Exitcustomer table with the
bank_churn table based on the ExitID column from Exitcustomer and the
Exited column from bank_churn. Finally, the results are grouped by the exit
category.

Output:

ExitCategory AvgCreditScore
Exit 645.35
Retain 651.85

While the difference in average credit scores between the two groups is
relatively small, it appears that customers who remain have a slightly higher
average credit score compared to those who have exited.

6. Which gender has a higher average estimated salary, and how

does it relate to the number of active accounts? (SQL)
SELECT g.GenderCategory,
AVG(c.EstimatedSalary) AS AvgEstimatedSalary,
 Avg_num_of_products
1.53

round(SUM(CASE WHEN b.IsActiveMember = 1 THEN 1 ELSE 0 END)) AS

ActiveAccounts
JOIN bank_churn b ON c.CustomerId = b.CustomerId
GROUP By g.GenderCategory;

FROM customerinfo c

JOIN gender g

ON c.GenderID = g.GenderID

It calculates the average estimated salary and the count of active accounts for
each gender category. It achieves this by joining three tables: customerinfo,
gender, and bank_churn. The gender table is linked to the customerinfo table
via the GenderID column, providing the gender category for each customer.
Meanwhile, the bank_churn table is connected to the customerinfo table
through the CustomerId column, allowing for the retrieval of churn-related
data. The query utilizes aggregate functions such as AVG and SUM, along with
conditional statements within the CASE expression, to compute the average
estimated salary and count the active accounts. Finally, the results are grouped
by gender category.

Output:

GenderCategory AvgEstimatedSalary ActiveAccounts

Female 100601.54 2284
Male 99664.57 2867
The data shows males have more active accounts compared to females. But
still, the AvgEstimatedSalary for females is higher than that of males, which
indicates that gender affects the AvgEstimatedSalary.

7. Segment the customers based on their credit score and identify

the segment with the highest exit rate. (SQL)
WITH CreditScoreSegments AS (
SELECT
CASE
WHEN CreditScore >= 800 AND CreditScore <=850 THEN 'Excellent'
WHEN CreditScore >= 740 AND CreditScore < 800 THEN 'Very Good'
WHEN CreditScore >= 670 AND CreditScore < 740 THEN 'Good'
WHEN CreditScore >= 580 AND CreditScore < 670 THEN 'Fair'
WHEN CreditScore >= 300 AND CreditScore < 580 THEN 'Poor'
ELSE 'Unknown'
END AS CreditScoreSegment,
CustomerId
FROM bank_churn)
SELECT CreditScoreSegment,
COUNT(CASE WHEN ExitCategory = 'Exit' THEN 1 END) AS ChurnedCustomers,
COUNT(*) AS TotalCustomers,
ROUND(COUNT(CASE WHEN ExitCategory = 'Exit' THEN 1 END) * 100.0 /
COUNT(*), 2) AS ExitRate
FROM CreditScoreSegments cs
JOIN bank_churn b ON cs.CustomerId = b.CustomerId
join exitcustomer ec on b.Exited=ec.ExitID
GROUP BY CreditScoreSegment
ORDER BY ExitRate DESC;
This SQL query segments customers based on their credit score into
predefined categories, such as Excellent, Very Good, Good, Fair, and Poor. It
calculates the count of churned customers, total customers, and exit rate for
each credit score segment. The "CreditScoreSegments" common table
expression (CTE) categorizes customers into segments based on their credit
score ranges. The main query then joins this CTE with the "bank_churn" table
to retrieve customer exit information and with the "exitcustomer" table to
obtain exit categories. Finally, it groups the results by credit score segment and
calculates the exit rate, ordering the output by exit rate in descending order.

Output:

CreditScoreSegment ExitedCustomers TotalCustomers ExitRate

Poor 520 2362 22.02
Very Good 252 1224 20.59
Fair 685 3331 20.56
Excellent 128 655 19.54
Good 452 2428 18.62
The data suggests that customers in the "Poor" credit score segment have the
highest exit rate at 22.02%. (The number of exited customer is maximum for
Fair segment but it has high total customers count.)

8. Find out which geographic region has the highest number of

active customers with a tenure greater than 5 years. (SQL)
SELECT g.GeographyLocation,
count(CASE WHEN b.IsActiveMember = 1 THEN 1 ELSE 0 END) AS
ActiveAccounts from geography g
join customerinfo c
on g.GeographyID=c.GeographyID
join bank_churn b
on c.CustomerId=b.CustomerId
where b.Tenure>5
group by g.GeographyLocation;
The SQL query identifies the geographic region with the highest number
of active customers who have a tenure greater than 5 years. It achieves this by
joining the geography, customerinfo, and bank_churn tables based on their
corresponding IDs. Within the WHERE clause, the query filters customers with
a tenure exceeding 5 years. Using a COUNT function with a conditional
statement, the query determines the count of active accounts in each
geographic region. The results are then grouped by geography location.

Output:

GeographyLocation ActiveAccounts
France 1575
Spain 796
Germany 800

9. What is the impact of having a credit card on customer churn,

based on the available data?
SELECT cc.Category AS CreditCardCategory,
COUNT(CASE WHEN ec.ExitCategory = 'Exit' THEN 1 END) AS
ChurnedCustomers,
COUNT(*) AS TotalCustomers,
ROUND(COUNT(CASE WHEN ec.ExitCategory = 'Exit' THEN 1 END) * 100.0 /
COUNT(*), 2) AS ChurnRate
FROM Creditcard cc
JOIN bank_churn ci ON cc.CreditID = ci.HasCrCard
JOIN Exitcustomer ec ON ci.Exited = ec.ExitID
GROUP BY cc.Category;
By joining the Creditcard, Customerinfo, and Exitcustomer tables, it
categorizes customers based on their credit card status and calculates the
churn rate for each category. Using conditional statements, the query
identifies churned customers within each credit card category and computes
the total number of customers. By dividing the number of churned customers
by the total and expressing the result as a percentage, it determines the churn
rate for each category.

Output:

CreditCardCategory ChurnedCustomer TotalCustomers ChurnRate

s
Credit card holder 1424 7055 20.18
Non credit card 613 2945 20.81
holder
In this case, both credit card holders and non-credit card holders exhibit
relatively similar churn rates, with credit card holders at 20.18% and non-credit
card holders at 20.81%. This suggests that, based on the available data, there is
no substantial difference in churn rates between customers who hold credit
cards and those who do not.

10. For customers who have exited, what is the most common
number of products they have used?
select b.NumOfProducts,
COUNT(CASE WHEN ec.ExitCategory = 'Exit' THEN 1 END) AS ExitedCustomers
from bank_churn b
join exitcustomer ec
on b.Exited=ec.ExitID
group by b.NumOfProducts
order by ExitedCustomers desc;
It retrieves data from the bank_churn and exitcustomer tables and
groups it by the number of products held by customers. Using a conditional
statement, it counts the number of customers who have churned within each
product category. The results are then ordered in descending order based on
the count of exited customers.

Output:

NumOfProducts ExitedCustomers
1 1409
2 348
3 220
4 60

It indicates that among exiting customers, the most common number of

products used is 1, with 1409 customers.

11. Examine the trend of customers joining over time and identify
any seasonal patterns (yearly or monthly). Prepare the data
through SQL and then visualize it.
SELECT
EXTRACT(YEAR FROM BankDOJ) AS JoinYear,
EXTRACT(MONTH FROM BankDOJ) AS JoinMonth,
COUNT(*) AS JoinCount
FROM customerinfo
GROUP BY JoinYear,JoinMonth
ORDER BY joinCount desc;

It examines the trend of customers joining over time by extracting the year and
month from the "BankDOJ" (Bank Date of Joining) column in the
"customerinfo" table. It counts the number of customers who joined in each
year-month combination and then groups the results by year and month.
Finally, it orders the results based on the join count in descending order.

Output:
 It appears that in the year 2019(470), December had the highest number
of customer joining’s, followed by November in 2018(368), December in
2017(334), and November in 2016(313). This analysis suggests that there is
seasonal pattern in customer acquisition, with higher joining’s typically
observed in the months of September, November, and December.

12. Analyze the relationship between the number of products and

the account balance for customers who have exited.

Based on the analysis, it shows that the majority of customer who

purchased 1 or 2 products exit more i.e., 5086 and 4590 for 1 and 2 products
respectively. In the above chart, a filter ExitID = 1 is used.
13. Identify any potential outliers in terms of balance among
customers who have remained with the bank.
SELECT
Exited,
COUNT(*) AS count_retained,
SUM(CASE WHEN Balance = 0 THEN 1 ELSE 0 END) AS count_zero_balance,
SUM(CASE WHEN Balance <> 0 THEN 1 ELSE 0 END) AS
count_nonzero_balance
FROM bank_churn
GROUP BY Exited;
Output:

Out of the 7963 retained customers, 3117 have a zero balance. This
indicates that some customers who have remained with the bank have fully
utilized their funds or closed their accounts. 4846 have a non-zero balance.
This suggests that the majority of customers who have remained with the bank
still have funds in their accounts.

14. How many different tables are given in the dataset, out of these
tables which table only consists of categorical variables?
There are seven different tables given in the dataset. Among these
tables, "Gender", "Geography", "ExitCustomer", and "ActiveCstomer" are likely
categorical variables.
15. Using SQL, write a query to find out the gender-wise average
income of males and females in each geography id. Also, rank the
gender according to the average value. (SQL)
SELECT c.GeographyID, g.GenderCategory,
AVG(c.EstimatedSalary) AS AvgIncome,
RANK() OVER (PARTITION BY c.GeographyID ORDER BY
AVG(c.EstimatedSalary) DESC) AS GenderRank
FROM customerinfo c
JOIN gender g ON c.GenderID = g.GenderID
GROUP BY c.GeographyID, g.GenderCategory
ORDER BY c.GeographyID, GenderRank;
It directly joins the customerinfo table with the gender table using the
JOIN clause based on the GenderID column and then it calculates the average
income (AVG(c.EstimatedSalary)) for each combination of GeographyID and
GenderCategory using the GROUP BY clause. The RANK() window function is
used to rank the genders within each geographic location based on their
average income. The PARTITION BY clause ensures that ranking is done
separately for each geographic location and gender category. Finally, the
results are ordered by GeographyID and GenderRank.

Output:

In two geographic region(Spain,Germany), the average income of

females is greater than that of male, it suggests that female income is higher
than male income across two locations.
16. Using SQL, write a query to find out the average tenure of the
people who have exited in each age bracket (18-30, 30-50, 50+).
SELECT CASE
WHEN Age BETWEEN 18 AND 30 THEN '18-30'
WHEN Age BETWEEN 31 AND 50 THEN '31-50'
ELSE '50+'
END AS AgeBracket,
AVG(Tenure) AS AvgTenure
FROM customerinfo ci
JOIN bank_churn bc ON ci.CustomerId = bc.CustomerId
JOIN exitcustomer ec ON bc.Exited = ec.ExitID
GROUP BY AgeBracket;
The customers are categorized into age brackets ('18-30', '31-50', and
'50+') using a CASE statement based on their age. The average tenure of
customers who have exited is then calculated by joining the necessary tables
(customerinfo, bank_churn, and exitcustomer) and utilizing the AVG() function.
Finally, the results are grouped by the AgeBracket column to obtain the
average tenure for each bracket.

Output:

AgeBracket AvgTenure (Years)

31-50 4.87
18-30 4.84
50+ 4.85

17. Is there any direct correlation between salary and the balance
of the customers? And is it different for people who have exited or
not?
 On average, customers who have exited the bank have a slightly higher
estimated salary compared to those who have remained. Customers who have
exited the bank have a notably higher average balance compared to those who
have remained. This suggests that customers with higher balances are more
likely to leave the bank.

Output:

18. Is there any correlation between the salary and the Credit score
of customers?
From the table it is evident that there is no relation between the Credit
Score of the customer and the Average of Estimated Salary. As the data
suggests, the lowest Credit score has the highest Average Estimated Salary.

Output:

19. Rank each bucket of credit score as per the number of

customers who have churned the bank.
SELECT CreditScoreBucket,
 COUNT(CASE WHEN ec.ExitCategory = 'Exit' THEN 1 END) AS
ChurnedCustomers,

RANK() OVER (ORDER BY COUNT(CASE WHEN ec.ExitCategory = 'Exit' THEN

1 END) DESC) AS Ranks

FROM (

SELECT CASE

WHEN CreditScore BETWEEN 800 AND 850 THEN 'Excellent'

WHEN CreditScore BETWEEN 740 AND 799 THEN 'Very Good'

WHEN CreditScore BETWEEN 670 AND 739 THEN 'Good'

WHEN CreditScore BETWEEN 580 AND 669 THEN 'Fair'

WHEN CreditScore BETWEEN 300 AND 579 THEN 'Poor'

END AS CreditScoreBucket,

CustomerId,Exited

FROM bank_churn

) AS ScoreBuckets

LEFT JOIN exitcustomer ec ON ScoreBuckets.Exited = ec.ExitID

GROUP BY CreditScoreBucket

ORDER BY Ranks;

This SQL query categorizes customers into credit score buckets ('Poor',
'Fair', 'Good', 'Excellent', 'Unknown') based on their credit scores. It then
calculates the number of churned customers within each bucket. Additionally,
it ranks the buckets according to the count of churned customers, with higher
ranks indicating more churn. The query utilizes a subquery to assign credit
score buckets to customers and then joins it with the exitcustomer table to
identify churned customers. Finally, it groups the results by credit score bucket
and orders them by rank in descending order.
Output:

The "Fair" credit score bucket is ranked first because it has the highest
number of churned customers compared to other credit score buckets, with
685 customers exiting the bank. It is noteworthy that a significant portion of
customers fall within the 600-700 credit score range, which corresponds to the
"Fair" credit score category.

20. According to the age buckets find the number of customers

who have a credit card. Also retrieve those buckets that have lesser
than average number of credit cards per bucket.
The given table contains the number of customers who have a credit
card across three age buckets. The Average number of credit cards per age
bucket is 2351.66 (7055/3). So, the age buckets 18-30 and 50 or above have
lesser values than the average number of credit cards.

Output:

21. Rank the Locations as per the number of people who have
churned the bank and average balance of the customers.
SELECT GeographyLocation,Num_Churned_Customers,Avg_Balance,
 RANK() OVER (ORDER BY Num_Churned_Customers DESC, Avg_Balance
DESC) AS Location_Rank
FROM(SELECT
geo.GeographyLocation,
COUNT(*) AS Num_Churned_Customers,
ROUND(AVG(bc.Balance),2) AS Avg_Balance
FROM bank_churn bc
JOIN CustomerInfo ci ON bc.CustomerId = ci.CustomerId
JOIN Geography geo ON ci.GeographyID = geo.GeographyID
WHERE bc.Exited = 1
GROUP BY geo.GeographyLocation) AS LocationStats;
This SQL query ranks the geographic locations based on the number of
customers who have churned the bank (Num_Churned_Customers) and their
average balance (Avg_Balance). It utilizes the RANK() function to assign a rank
to each location, ordering them by the number of churned customers in
descending order, and then by the average balance in descending order.

Output:

Germany has the highest number of churned customers (814) and the
highest average balance among them (120,361.08), thus it is ranked first.

22. As we can see that the “CustomerInfo” table has the

CustomerID and Surname, now if we have to join it with a table
where the primary key is also a combination of CustomerID and
Surname, come up with a column where the format is
“CustomerID_Surname”.
SELECT CONCAT(ci.CustomerID, '_', ci.Surname) AS CustomerID_Surname
FROM CustomerInfo ci
JOIN OtherTable ot ON ci.CustomerID = ot.CustomerID AND ci.Surname =
ot.Surname;
This query joins the "CustomerInfo" table with the "bankchurn" based on the
CustomerID column. The CONCAT function is then used to concatenate the
CustomerID and Surname columns with an underscore (_) separator to create
the new column "CustomerID_Surname".

Output:

Each row represents a unique customer identified by their CustomerID

and Surname. For example, "15634602_Hargrave" indicates a customer with a
CustomerID of 15634602.

23. Without using “Join”, can we get the “ExitCategory” from

ExitCustomers table to Bank_Churn table? If yes do this using SQL.
Yes, we get the ExitCategory” from ExitCustomers table to Bank_Churn
table by using subquery
SELECT *,(SELECT ExitCategory FROM exitcustomer WHERE ExitID = bc.Exited)
AS ExitCategory FROM bank_churn bc;
This SQL query selects all columns from the table bank_churn.
Additionally, it includes a subquery that retrieves the ExitCategory from the
exitcustomer table based on a condition matching ExitID with Exited from the
bank_churn table. The result includes an extra column named ExitCategory.
Output:

24. Were there any missing values in the data, using which tool did
you replace them and what are the ways to handle them?
The dataset does not contain any missing values. The only change made to the
dataset was converting the datatype of the BankDOJ column to a date
datatype.

25. Write the query to get the customer IDs, their last name, and
whether they are active or not for the customers whose surname
ends with “on”.
SELECT ci.CustomerId,ci.Surname,
MAX(ac.ActiveCategory) AS ActiveCategory
FROM customerinfo ci
JOIN bank_churn b ON ci.CustomerId = b.CustomerId
JOIN activecustomer ac ON b.IsActiveMember = ac.ActiveID
WHERE ci.Surname LIKE '%on'
GROUP BY ci.CustomerId,ci.Surname;
This SQL query selects unique combinations of CustomerId and Surname from
the customerinfo table where the Surname contains "on". It retrieves the
maximum ActiveCategory associated with each combination from the
activecustomer table.

Sample Output:

This query groups results by CustomerId and Surname, selecting the highest
ActiveCategory for each group, ensuring unique CustomerId entries in the
output.

Subjective Questions:
1. Customer Behavior Analysis: What patterns can be observed in
the spending habits of long-term customers compared to new
customers, and what might these patterns suggest about customer
loyalty?

Long term customers who have retained in the bank have purchased
more number of products when compared to new customers.
It is to be noted that the Number of customers and their Activeness are
more in this category.
The reason behind the Low average balance among other categories
might be because they have purchased more products.

This indicates that long-term customers have developed deeper

relationships with the bank over time and are more likely to utilize multiple
products and services, which suggests a deeper level of engagement and
loyalty to the bank.

2. Product Affinity Study: Which bank products or services are most

commonly used together, and how might this influence cross-
selling strategies?

There are high numbers of customers with only one or two products.
This suggests that there is a strong association between the number of
products a customer holds and their possession with the bank. Targeted
marketing campaigns and personalized recommendations can be tailored
based on customers' existing product holdings to encourage them to expand
their relationships with the bank.

3. Geographic Market Trends: How do economic indicators in

different geographic regions correlate with the number of active
accounts and customer churn rates?

Germany has the highest churn rate at 32.44%, followed by Spain at

16.67%, and France at 16.15%. Churn rates vary across geographic
regions, indicating potential differences in economic conditions,
customer behavior, or competitive landscapes.
 France and Spain have relatively similar counts of active accounts
despite differences in churn rates.

In regions with high churn rates, bank may need to focus on improving
customer retention strategies, such as enhancing customer service, offering
competitive rates, or introducing loyalty programs.

Similarly, in regions with lower counts of active accounts, bank may need to
invest in marketing efforts, product innovation, or partnerships to attract new
customers and expand their market share.

4. Risk Management Assessment: Based on customer profiles,

which demographic segments appear to pose the highest financial
risk to the bank, and why?

Customers aged 50 or above demonstrate a significantly higher churn

rate compared to younger age groups, indicating that they pose a higher
financial risk to the bank.
Despite having a higher average balance, older customers are more
likely to churn, suggesting that factors beyond account balance influence
their decision to leave the bank.
It suggests that older customers warrant special attention in risk
management strategies to mitigate potential losses associated with
churn.
5. Customer Tenure Value Forecast: How would you use the
available data to model and predict the lifetime (tenure) value in
the bank of different customer segments?

There are noticeable differences in the sum of estimated salaries across

France, Germany, and Spain. France and Germany generally have higher
total estimated salaries compared to Spain, indicating potential
economic disparities between these regions.
While France and Germany have higher total estimated salaries, the
tenure patterns vary across different geographic locations. France and
Germany show a similar trend of increasing estimated salaries up to a
certain tenure (around 5 years) before declining for longer tenures. In
contrast, Spain exhibits a similar trend but with lower estimated salaries
overall.
The analysis of tenure distribution suggests that customers in France and
Germany tend to have longer relationships with the bank, as indicated
by their tenure extending up to 7 years. This may imply higher customer
loyalty or satisfaction levels in these regions compared to Spain.

6. Marketing Campaign Effectiveness: How could you assess the

impact of marketing campaigns on customer retention and
acquisition within the dataset? What extra information would you
need to solve this?
From the table, it is observed that the credit score segment of Excellent and
Very good has the least count of exited customers which shows that the
improvement of credit score impacts in the retention rate of the customers. By
conducting Marketing Campaigns the Credit score of the customer will
increase which will directly impact on the retention rates.

To effectively assess the impact of marketing campaigns on customer

retention and acquisition within the dataset, additional information that would
be helpful includes:

Data on the customer journey from initial awareness to conversion,

including touchpoints and interactions with marketing channels.
Information on the cost associated with each marketing campaign,
including ad spend, creative production costs, and other expenses.
Feedback from customers regarding their experience with marketing
campaigns, perceived value, and factors influencing their decision-
making.
Insights into competitor marketing strategies, messaging, and market
positioning to understand the competitive landscape and potential
impact on campaign effectiveness.

By combining these additional information, a comprehensive analysis can

be conducted to evaluate the effectiveness of marketing campaigns on
customer retention and acquisition.

7. Customer Exit Reasons Exploration: Can you identify common

characteristics or trends among customers who have exited that
could explain their reasons for leaving?
  The churn rate is notably higher for customers aged 50 or above
compared to other age groups. This indicates that older customers may
be more likely to leave the bank.
 Customers who exited the bank have a significant total balance and
estimated salary compared to others. This suggests that financial factors
may play a role in their decision to leave.

Some of the possible reasons for leaving:

Dissatisfaction with banking services or customer experiences.

Inadequate credit card benefits or rewards.
Financial instability or changes in personal circumstances.
Better offers or services from competitors.
Lack of tailored products or solutions for older customers.

8. Are 'Tenure', 'NumOfProducts', 'IsActiveMember', and

'EstimatedSalary' important for predicting if a customer will leave
the bank?

The sum of tenure for retained customers (0) is significantly higher

compared to exited customers (1). This suggests that customers who
have been with the bank for longer periods are less likely to leave.
 The sum of the number of products for retained customers (0) is higher
compared to exited customers (1) across all age groups. This indicates
that customers who use more products/services offered by the bank are
less likely to churn.
The sum of IsActiveMember for retained customers (0) is higher
compared to exited customers (1) across all age groups. This suggests
that customers who are active members are less likely to churn.
The sum of estimated salary for retained customers (0) is significantly
higher compared to exited customers (1) across all age groups. This
implies that customers with higher estimated salaries are less likely to
leave the bank.
Based on the provided data, 'Tenure', 'NumOfProducts',
'IsActiveMember', and 'EstimatedSalary' appear to be important factors
for predicting if a customer will leave the bank. Customers with longer
tenure, higher usage of products/services, active membership status,
and higher estimated salaries are less likely to churn.

9. Utilize SQL queries to segment customers based on

demographics and account details.
SELECT g.GenderCategory,

CASE

WHEN c.Age BETWEEN 18 AND 30 THEN '18-30'

WHEN c.Age BETWEEN 31 AND 50 THEN '31-50'

ELSE '50 or above'

END AS AgeBin,

COUNT(*) AS CustomerCount,

AVG(bc.balance) AS AverageBalance

FROM customerinfo c

JOIN bank_churn bc ON c.CustomerId = bc.CustomerId

JOIN activecustomer ac ON bc.IsActiveMember = ac.ActiveID

JOIN Gender g ON c.GenderId = g.GenderId

JOIN creditcard cc ON bc.HasCrCard = cc.CreditID

JOIN exitcustomer ec ON bc.Exited = ec.ExitID

GROUP BY g.GenderCategory,

CASE

WHEN c.Age BETWEEN 18 AND 30 THEN '18-30'

WHEN c.Age BETWEEN 31 AND 50 THEN '31-50'

ELSE '50 or above'

END;

SELECT

-- Demographic Segmentation

CASE

WHEN ci.Age BETWEEN 18 AND 30 THEN '18-30'

WHEN ci.Age BETWEEN 31 AND 50 THEN '31-50'

WHEN ci.Age > 50 THEN '50+'

ELSE 'Unknown'

END AS Age_Group,

g.GeographyLocation,

gi.GenderCategory AS Gender,

-- Account Details Segmentation

CASE

WHEN bc.Balance < 10000 THEN 'Low Balance'

 WHEN bc.Balance >= 10000 AND bc.Balance < 50000 THEN 'Medium
Balance'

WHEN bc.Balance >= 50000 THEN 'High Balance'

ELSE 'Unknown'

END AS Balance_Category,

cc.Category AS Credit_Card_Status

FROM customerinfo ci

JOIN bank_churn bc ON ci.CustomerId = bc.CustomerId

JOIN geography g ON ci.GeographyID = g.GeographyID

JOIN gender gi ON ci.GenderID = gi.GenderID

JOIN creditcard cc ON bc.HasCrCard = cc.CreditID;

The SQL query segments customers based on demographic attributes

like age group, geography, and gender, as well as account details such as
balance category and credit card status. It utilizes CASE statements to
categorize customers into appropriate groups and performs joins across
multiple tables to gather necessary information. The resulting output provides
a segmented view of customers, facilitating analysis of their characteristics and
behaviors across different segments.
10. How can we create a conditional formatting setup to visually
highlight customers at risk of churn and to evaluate the impact of
credit card rewards on customer retention?

In Power Bi table visualization is selected to apply conditional

formatting. Right-click on the field or measure to format and choose
"Conditional formatting" from the context menu.
In the Conditional formatting pane, the rules are defined based on Chun
rate and Credit card criteria’s.
Specify the formatting options for each rule. Different formatting styles
are chosen such as colors and Data bar is used to represent different
levels of risk or impact.
Once the rules are defined in the formatting options, click "Apply" or
"OK" to apply the conditional formatting to the visualization.

The visualization enables comparison across different age groups to

understand churn rates, credit card ownership, and credit score
distributions. It highlights potential trends or correlations between age,
credit card usage, credit scores, and churn rates, providing valuable
 insights for decision-making and targeting specific customer segments
for retention efforts or marketing campaigns.
From the table chart, it is evident that the agebin 50+ has the highest
churn rate which implies that they are likely prone to leave the bank.
Also, this age groups has the least number of credit cards.
Similarly, customers having credit cards are less likely to leave the bank
which shows the impact of credit card rewards in the customers.

11. What is the current churn rate per year and overall as well in
the bank? Can you suggest some insights to the bank about which
kind of customers are more likely to churn and what different
strategies can be used to decrease the churn rate?
 The churn rates vary across different age groups each year. Generally,
older customers (50 or above) consistently exhibit higher churn rates
compared to younger age groups (18-30 and 30-50).While there are
fluctuations in churn rates from year to year, there isn't a significant upward or
downward trend in any specific age group.

Strategies to Decrease Churn Rate:

The bank can conduct marketing and engagement campaigns tailored to

the unique needs and preferences of different age groups, such as
offering retirement planning services and personalized financial advice
for older customers.
Specialized customer support services, including dedicated helplines for
senior citizens or personalized assistance for complex financial
transactions, can be provided to improve satisfaction and retention
among older customers.
Conducting financial literacy workshops and educational seminars
targeted at different age groups can empower customers with
knowledge and skills to manage their finances effectively, reducing the
likelihood of churn.
Implementing loyalty programs and retention incentives customized for
each age group, including rewards, discounts, and exclusive benefits, can
enhance customer loyalty and decrease churn rates.
 Establishing feedback mechanisms to gather insights directly from
customers about their needs, preferences, and pain points can inform
continuous refinement and optimization of retention strategies for each
age group.

12. Create a dashboard incorporating all the KPIs and visualization-

related metrics. Use a slicer in order to assist in selection in the
dashboard.
13. How would you approach this problem, if the objective and
subjective questions weren't given?
If the objective and subjective questions weren't provided, I would
approach the problem by :

First understanding the dataset thoroughly, including its structure,

variables, and relationships between them.
I would perform exploratory data analysis to find out the uncover
patterns, trends, and insights within the data.
I would involve visualizing data through charts and graphs, and
identifying any correlations or relationships between variables.
I would identify customer churn rate, customer retention rate, average
account balance, demographics distribution, and product usage metrics.
I would then use SQL queries to calculate these metrics and present
them in a meaningful way.
After that, I would conduct deeper analysis to answer specific questions
or address particular objectives, such as identifying factors influencing
customer churn, understanding customer segmentation based on
demographics or behavior.
14. In the “Bank_Churn” table how can you modify the name of
the “HasCrCard” column to “Has_creditcard”?
In Power BI, you can rename columns in the query editor. Here's how you can
modify the name of the "HasCrCard" column to "Has_creditcard":

Open your Power BI desktop file.

Go to the "Home" tab and click on "Transform Data" to open the query editor.

In the query editor, find the table "Bank_Churn" from the list of queries on the
left-hand side.

Select the "Bank_Churn" query to display its columns and data.

Right-click on the "HasCrCard" column header and select "Rename".

Replace "HasCrCard" with "Has_creditcard" and press Enter to confirm.

Once you've renamed the column, click on "Close & Apply" to save your
changes and close the query editor.