Distribution analysis (box plots) in Tableau - Real Business Scenario

Scenario Mode

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

📋 Request: Your manager wants to understand the distribution of sales amounts across different product categories to identify variability and outliers.

📊 Data: You have sales transaction data including Product Category, Sales Amount, and Transaction Date.

🎯 Deliverable: Create a dashboard with box plots showing the distribution of sales amounts for each product category.

Sample Data

Transaction ID	Product Category	Sales Amount	Transaction Date
1001	Electronics	250	2024-05-01
1002	Electronics	300	2024-05-02
1003	Furniture	450	2024-05-01
1004	Furniture	700	2024-05-03
1005	Clothing	80	2024-05-02
1006	Clothing	120	2024-05-04
1007	Electronics	150	2024-05-05
1008	Furniture	600	2024-05-05
1009	Clothing	200	2024-05-06
1010	Electronics	400	2024-05-06

1

Step 1: Connect Tableau to the sales transaction data source.

Use the provided data table with columns: Transaction ID, Product Category, Sales Amount, Transaction Date.

Expected Result

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

2

Step 2: Create a new worksheet for the box plot.

Drag 'Product Category' to Columns shelf. Drag 'Sales Amount' to Rows shelf.

Expected Result

A bar chart appears showing total sales amount by product category.

3

Step 3: Change the mark type to 'Box-and-Whisker Plot'.

Click on the Marks card dropdown and select 'Box-and-Whisker Plot'.

Expected Result

Box plots appear for each product category showing distribution of sales amounts.

4

Step 4: Add tooltips to show minimum, first quartile, median, third quartile, and maximum sales amounts.

Ensure default box plot tooltips are enabled in Tableau.

Expected Result

Hovering over each box plot shows detailed distribution statistics.

5

Step 5: Format the visualization for clarity and accessibility.

Add axis titles: 'Product Category' (X-axis), 'Sales Amount' (Y-axis). Use color to differentiate categories if desired. Ensure color contrast is sufficient.

Expected Result

Box plots are clearly labeled and easy to interpret.

6

Step 6: Create a dashboard and add the box plot worksheet.

Drag the worksheet onto a new dashboard. Add a title: 'Sales Amount Distribution by Product Category'.

Expected Result

Dashboard displays the box plots with a clear title.

Final Result

Dashboard: Sales Amount Distribution by Product Category

+----------------+----------------+----------------+
| Electronics    | Furniture      | Clothing       |
|  ┌─────┐       |  ┌─────┐       |  ┌─────┐       |
|  │  ■  │       |  │  ■  │       |  │  ■  │       |
|  │■■■■■│       |  │■■■■■│       |  │■■■■■│       |
|  └─────┘       |  └─────┘       |  └─────┘       |
|  Min 150       |  Min 450       |  Min 80        |
|  Median 275    |  Median 600    |  Median 120    |
|  Max 400       |  Max 700       |  Max 200       |
+----------------+----------------+----------------+

✓Electronics sales have moderate variability with some outliers at higher sales amounts.

✓Furniture sales show higher sales amounts and wider distribution, indicating more variability.

✓Clothing sales have lower sales amounts and less variability compared to other categories.

Bonus Challenge

Add a filter to the dashboard to allow users to select a date range and see how sales distribution changes over time.

Show Hint

Use Tableau's date filter feature and apply it to the data source or worksheet to dynamically update the box plots.

Practice

1. What does a box plot primarily show in data visualization?

easy

A. The total count of data points

B. The spread of data including median, quartiles, and outliers

C. Only the average value of the data

D. The correlation between two variables

Solution

1. Step 1: Understand box plot components

   A box plot displays the median, quartiles (Q1 and Q3), and any outliers in the data.

2. Step 2: Compare options to definition

   Only the spread of data including median, quartiles, and outliers correctly describes these features; others describe different charts or statistics.

3. Final Answer:

   The spread of data including median, quartiles, and outliers -> Option B

4. Quick Check:

   Box plot = median + quartiles + outliers [OK]

Hint: Box plots show data spread, not just averages [OK]

Common Mistakes:

• Confusing box plots with bar charts
• Thinking box plots show only averages
• Ignoring outliers in box plots

2. Which Tableau feature helps you quickly create a box plot?

easy

A. Using the 'Show Me' panel and selecting box plot

B. Writing a custom SQL query

C. Using the Data Interpreter tool

D. Applying a filter on the data source

Solution

1. Step 1: Identify Tableau tools for visualization

   Tableau's 'Show Me' panel offers quick chart types including box plots.

2. Step 2: Evaluate other options

   Custom SQL, Data Interpreter, and filters do not create visualizations directly.

3. Final Answer:

   Using the 'Show Me' panel and selecting box plot -> Option A

4. Quick Check:

   'Show Me' = quick box plot creation [OK]

Hint: Use 'Show Me' to pick box plot instantly [OK]

Common Mistakes:

• Confusing data preparation tools with visualization tools
• Trying to write SQL for box plots in Tableau
• Thinking filters create charts

3. Given a box plot in Tableau showing sales by region, what does the line inside the box represent?

medium

A. The average sales value

B. The maximum sales value

C. The median sales value

D. The total sales sum

Solution

1. Step 1: Recall box plot components

   The line inside the box is the median, which divides data into two equal halves.

2. Step 2: Differentiate median from average and extremes

   The average is the mean, not shown by the line; max is the whisker end; total sum is not shown.

3. Final Answer:

   The median sales value -> Option C

4. Quick Check:

   Box plot line = median [OK]

Hint: Median is the line inside the box, not average [OK]

Common Mistakes:

• Confusing median with average
• Thinking the line shows max or total
• Ignoring box plot whiskers

4. You created a box plot in Tableau but the whiskers are missing. What is the most likely cause?

medium

A. You forgot to add a measure to the Rows shelf

B. The data contains no outliers or extreme values

C. The data source is not connected

D. You applied a filter that removed all data

Solution

1. Step 1: Understand box plot requirements in Tableau

   Box plots need a measure on Rows or Columns to calculate quartiles and whiskers.

2. Step 2: Analyze each option

   No whiskers usually mean no measure is assigned; data connection or filters removing all data would prevent any plot.

3. Final Answer:

   You forgot to add a measure to the Rows shelf -> Option A

4. Quick Check:

   Missing whiskers = no measure assigned [OK]

Hint: Always add a measure to Rows or Columns for box plots [OK]

Common Mistakes:

• Assuming no whiskers means no outliers only
• Ignoring the need for measures in visualization
• Not checking data connection status

5. You want to compare sales distributions across multiple product categories using box plots in Tableau. Which approach best ensures clear comparison?

hard

A. Filter data to one category at a time and view box plots

B. Create separate worksheets for each category and compare manually

C. Use a pie chart to show sales proportions per category

D. Place product category on Columns and sales measure on Rows, then select box plot from 'Show Me'

Solution

1. Step 1: Set up box plot for multiple categories

   Placing product category on Columns and sales on Rows allows Tableau to create side-by-side box plots for comparison.

2. Step 2: Evaluate other options

   Separate worksheets or filtering one category at a time prevents direct visual comparison; pie charts do not show distribution.

3. Final Answer:

   Place product category on Columns and sales measure on Rows, then select box plot from 'Show Me' -> Option D

4. Quick Check:

   Side-by-side box plots = best for category comparison [OK]

Hint: Use Columns for categories and Rows for measure to compare box plots [OK]

Common Mistakes:

• Using pie charts instead of box plots for distribution
• Comparing categories in separate sheets
• Filtering to one category losing comparison context