Clustering in Tableau - Real Business Scenario

Scenario Mode

👤 Your Role: You are a sales analyst at a retail company.

📋 Request: Your manager wants to identify groups of customers with similar buying behaviors to target marketing campaigns better.

📊 Data: You have customer purchase data including Customer ID, Total Purchase Amount, Number of Transactions, and Average Basket Size.

🎯 Deliverable: Create a Tableau dashboard that uses clustering to segment customers into meaningful groups and visualize these clusters.

Sample Data

Customer ID	Total Purchase Amount	Number of Transactions	Average Basket Size
C001	1200	15	80
C002	300	5	60
C003	1500	20	75
C004	700	10	70
C005	200	3	67
C006	1800	25	72
C007	400	7	57
C008	1300	18	72
C009	600	9	66
C010	250	4	62

1

Step 1: Connect your customer purchase data to Tableau.

Use 'Connect to Data' and select your data source containing Customer ID, Total Purchase Amount, Number of Transactions, and Average Basket Size.

Expected Result

Data is loaded and visible in Tableau's Data pane.

2

Step 2: Create a scatter plot to visualize customer purchase behavior.

Drag 'Number of Transactions' to Columns and 'Total Purchase Amount' to Rows.

Expected Result

Scatter plot shows customers distributed by transactions and purchase amount.

3

Step 3: Add clustering to segment customers.

Click 'Analytics' pane, drag 'Cluster' onto the scatter plot, and let Tableau suggest clusters.

Expected Result

Customers are grouped into clusters with different colors.

4

Step 4: Adjust cluster settings to use 'Average Basket Size' as an additional variable.

In the cluster configuration, add 'Average Basket Size' as a variable for clustering.

Expected Result

Clusters update to reflect similarity across three variables.

5

Step 5: Create a dashboard to display the cluster visualization and customer details.

Add the scatter plot with clusters and a customer details table to a new dashboard.

Expected Result

Dashboard shows clusters visually and lists customers with their cluster assignment.

Final Result

Clusters Dashboard

+---------------------------------------+
| Scatter Plot: Number of Transactions  |
| vs Total Purchase Amount (colored by |
| cluster)                             |
|                                       |
|   *   *   *   *   *   *   *   *   *   |
|   *   *   *   *   *   *   *   *   *   |
|                                       |
+---------------------------------------+
| Customer Details Table                 |
| Customer ID | Cluster | Total Purchase|
| C001       | 1       | 1200          |
| C002       | 2       | 300           |
| ...                                   |
+---------------------------------------+

✓Customers are grouped into 3 clusters based on purchase amount, transaction count, and basket size.

✓Cluster 1 includes high-value customers with many transactions and large basket sizes.

✓Cluster 2 includes low-value customers with fewer transactions.

✓Cluster 3 includes medium-value customers with moderate transactions and basket sizes.

Bonus Challenge

Add a filter to the dashboard to allow users to select clusters and see only those customers.

Show Hint

Use Tableau's filter action on the cluster field to enable interactive filtering.

Practice

1.

What is the main purpose of clustering in Tableau?

easy

A. To create bar charts

B. To sort data alphabetically

C. To group similar data points automatically

D. To filter data by date

Solution

1. Step 1: Understand clustering concept

   Clustering groups data points that share similar characteristics without manual labeling.

2. Step 2: Identify Tableau's clustering use

   Tableau uses clustering to find natural groups in data automatically.

3. Final Answer:

   To group similar data points automatically -> Option C

4. Quick Check:

   Clustering = Group similar data [OK]

Hint: Clustering means grouping similar items automatically [OK]

Common Mistakes:

• Confusing clustering with sorting
• Thinking clustering creates charts
• Assuming clustering filters data

2.

Which step correctly adds clusters in Tableau?

1. Drag fields to Rows and Columns
2. Open Analytics pane
3. Drag 'Cluster' onto the view
4. Adjust cluster count

easy

A. Use the 'Show Me' panel to add clusters

B. Drag 'Cluster' from Data pane to view

C. Right-click view and select 'Add Cluster'

D. Drag 'Cluster' from Analytics pane to view

Solution

1. Step 1: Identify where 'Cluster' is located

   In Tableau, 'Cluster' is found in the Analytics pane, not Data pane or menus.

2. Step 2: Confirm correct action to add clusters

   You drag 'Cluster' from Analytics pane onto the view to create clusters.

3. Final Answer:

   Drag 'Cluster' from Analytics pane to view -> Option D

4. Quick Check:

   Clusters come from Analytics pane [OK]

Hint: Clusters come from Analytics pane, not Data pane [OK]

Common Mistakes:

• Trying to drag 'Cluster' from Data pane
• Looking for 'Add Cluster' in right-click menu
• Using 'Show Me' panel for clusters

3.

Given this Tableau clustering setup, what is the expected result?

- Data: Customers with sales and profit
- Clusters: 3 groups based on sales and profit
- View: Scatter plot of Sales vs Profit with clusters colored

What will the clusters show?

medium

A. Three groups of customers with similar sales and profit patterns

B. Three groups sorted by customer name alphabetically

C. Three groups based on sales only, ignoring profit

D. Random groups unrelated to sales or profit

Solution

1. Step 1: Understand clustering variables

   Clusters use sales and profit to group customers with similar values in both.

2. Step 2: Interpret cluster meaning on scatter plot

   Each cluster groups customers close in sales and profit, shown by colors.

3. Final Answer:

   Three groups of customers with similar sales and profit patterns -> Option A

4. Quick Check:

   Clusters group by sales and profit [OK]

Hint: Clusters group by all selected fields, not just one [OK]

Common Mistakes:

• Assuming clusters sort alphabetically
• Thinking clusters use only one variable
• Believing clusters are random

4.

What is wrong with this clustering setup in Tableau?

- Added 'Cluster' from Analytics pane
- Selected only one field: 'Customer ID'
- Result: All data points in one cluster

medium

A. Cluster count must be set to zero to work

B. Using only one unique ID field prevents meaningful clusters

C. Clusters require at least three fields to function

D. Cluster feature is not available for Customer ID field

Solution

1. Step 1: Analyze field choice for clustering

   Using a unique identifier like Customer ID does not group similar data; each ID is unique.

2. Step 2: Understand cluster result

   Since all IDs are unique, Tableau places all points in one cluster by default.

3. Final Answer:

   Using only one unique ID field prevents meaningful clusters -> Option B

4. Quick Check:

   Unique IDs don't create clusters [OK]

Hint: Avoid unique ID fields alone for clustering [OK]

Common Mistakes:

• Setting cluster count to zero
• Thinking clusters need 3+ fields
• Believing Customer ID can't be used at all

5.

You want to segment customers into 4 clusters using sales, profit, and discount in Tableau. Which approach is best?

1. Create a scatter plot with sales and profit
2. Add discount as a cluster variable
3. Drag 'Cluster' from Analytics pane to view
4. Set number of clusters to 4

What should you do to ensure meaningful clusters?

hard

A. Include all three fields (sales, profit, discount) in clustering and set clusters to 4

B. Use only sales and profit, ignore discount, and set clusters to 4

C. Set clusters to 2 for better separation with three fields

D. Create separate clusters for each field individually

Solution

1. Step 1: Include all relevant fields for clustering

   Using sales, profit, and discount together helps find groups based on all factors.

2. Step 2: Set cluster count as desired

   Setting clusters to 4 matches the goal of segmenting into four groups.

3. Final Answer:

   Include all three fields (sales, profit, discount) in clustering and set clusters to 4 -> Option A

4. Quick Check:

   Use all fields + 4 clusters [OK]

Hint: Use all relevant fields and set cluster count as needed [OK]

Common Mistakes:

• Ignoring discount field
• Choosing wrong cluster count
• Creating separate clusters per field