Why advanced analytics uncovers hidden patterns in Tableau - Formula Trace Breakdown

Sample Data

Sample sales data showing customer purchases with amounts and dates.

Cell	Value
A1	CustomerID
B1	PurchaseAmount
C1	PurchaseDate
A2	C001
B2	120
C2	2024-01-05
A3	C002
B3	80
C3	2024-01-06
A4	C001
B4	150
C4	2024-02-10
A5	C003
B5	200
C5	2024-02-15
A6	C002
B6	90
C6	2024-03-01

Formula Trace

WINDOW_AVG(SUM([PurchaseAmount]), 0, 2)

Step 1: SUM([PurchaseAmount]) for each row

Step 2: WINDOW_AVG(SUM([PurchaseAmount]), 0, 2) at row 1

Step 3: WINDOW_AVG(SUM([PurchaseAmount]), 0, 2) at row 2

Step 4: WINDOW_AVG(SUM([PurchaseAmount]), 0, 2) at row 3

Step 5: WINDOW_AVG(SUM([PurchaseAmount]), 0, 2) at row 4

Step 6: WINDOW_AVG(SUM([PurchaseAmount]), 0, 2) at row 5

Cell Reference Map

     A           B              C
1 |CustomerID |PurchaseAmount|PurchaseDate
2 |   C001    |    120      | 2024-01-05
3 |   C002    |     80      | 2024-01-06
4 |   C001    |    150      | 2024-02-10
5 |   C003    |    200      | 2024-02-15
6 |   C002    |     90      | 2024-03-01

Arrows: WINDOW_AVG uses SUM([PurchaseAmount]) from current and next 2 rows

The formula uses PurchaseAmount values from the current row and the next two rows to calculate a moving average.

Result

     A           B              C             D
1 |CustomerID |PurchaseAmount|PurchaseDate|MovingAvg
2 |   C001    |    120      | 2024-01-05 | 116.67
3 |   C002    |     80      | 2024-01-06 | 143.33
4 |   C001    |    150      | 2024-02-10 | 146.67
5 |   C003    |    200      | 2024-02-15 | 145.00
6 |   C002    |     90      | 2024-03-01 |  90.00

The MovingAvg column shows the average purchase amount over the current and next two purchases, revealing trends and smoothing out spikes.

Sheet Trace Quiz - 3 Questions

What does the WINDOW_AVG function calculate in this example?

A The average purchase amount for each customer

B The average of the current and next two purchase amounts

C The total sum of all purchase amounts

D The maximum purchase amount in the dataset

Key Result

WINDOW_AVG calculates the average of a measure over a sliding window of rows.

Practice

1. What is the main benefit of using advanced analytics in Tableau?

easy

A. It replaces the need for any manual data analysis.

B. It automatically cleans all data errors without user input.

C. It helps uncover hidden patterns in data that are not obvious.

D. It only creates simple bar charts and pie charts.

Solution

1. Step 1: Understand the purpose of advanced analytics

   Advanced analytics is designed to find insights that are not easily seen by simple observation.

2. Step 2: Identify Tableau's role in advanced analytics

   Tableau provides tools like clustering and forecasting to reveal hidden data patterns.

3. Final Answer:

   It helps uncover hidden patterns in data that are not obvious. -> Option C

4. Quick Check:

   Advanced analytics = uncover hidden patterns [OK]

Hint: Advanced analytics reveals what simple views miss [OK]

Common Mistakes:

• Thinking it automatically fixes data errors
• Believing it removes need for manual analysis
• Assuming it only makes basic charts

2. Which Tableau feature is correctly used to group similar data points automatically?

easy

A. Clustering

B. Forecasting

C. Trend Lines

D. Highlight Table

Solution

1. Step 1: Identify Tableau features for grouping

   Clustering groups similar data points based on patterns automatically.

2. Step 2: Differentiate from other features

   Trend lines show trends, forecasting predicts future values, highlight tables emphasize data but do not group.

3. Final Answer:

   Clustering -> Option A

4. Quick Check:

   Grouping similar data = Clustering [OK]

Hint: Clustering groups data automatically in Tableau [OK]

Common Mistakes:

• Confusing trend lines with grouping
• Thinking forecasting groups data
• Assuming highlight tables group data

3. Given a Tableau scatter plot with clustering applied, what is the expected output?

medium

A. Data points are hidden except for the largest values.

B. All data points are shown in a single color without grouping.

C. The plot shows a line predicting future values.

D. Data points are colored and grouped into clusters based on similarity.

Solution

1. Step 1: Understand clustering effect on scatter plot

   Clustering colors and groups data points that share similar characteristics.

2. Step 2: Eliminate incorrect outputs

   Single color means no clustering, line prediction is forecasting, hiding points is filtering, not clustering.

3. Final Answer:

   Data points are colored and grouped into clusters based on similarity. -> Option D

4. Quick Check:

   Clustering output = grouped colored points [OK]

Hint: Clusters color and group similar points [OK]

Common Mistakes:

• Confusing clustering with forecasting
• Expecting no color changes
• Thinking clustering hides data points

4. You applied forecasting in Tableau but the forecast line does not appear. What is the most likely issue?

medium

A. Clustering was applied instead of forecasting.

B. The data does not have a time or date field to base the forecast on.

C. The data source is not connected to Tableau.

D. The worksheet is filtered to show only one data point.

Solution

1. Step 1: Check forecasting requirements

   Forecasting needs a time or date field to predict future values.

2. Step 2: Identify why forecast line is missing

   Without a time/date field, Tableau cannot generate a forecast line.

3. Final Answer:

   The data does not have a time or date field to base the forecast on. -> Option B

4. Quick Check:

   Forecast needs time/date field [OK]

Hint: Forecast requires time or date data [OK]

Common Mistakes:

• Confusing clustering with forecasting
• Assuming data connection issues cause missing forecast
• Not checking filters before forecasting

5. A sales manager wants to use Tableau to find hidden customer segments based on purchase behavior and predict future sales trends. Which combination of Tableau features should they use?

hard

A. Clustering to find segments and Forecasting to predict sales trends.

B. Highlight Table to find segments and Trend Lines to predict sales.

C. Filters to find segments and Pie Charts to predict sales.

D. Parameters to find segments and Maps to predict sales.

Solution

1. Step 1: Identify feature for finding customer segments

   Clustering groups customers by similar purchase behavior, revealing hidden segments.

2. Step 2: Identify feature for predicting future sales

   Forecasting uses historical data to predict future sales trends.

3. Step 3: Confirm the correct combination

   Clustering and Forecasting together address both segmentation and prediction needs.

4. Final Answer:

   Clustering to find segments and Forecasting to predict sales trends. -> Option A

5. Quick Check:

   Segments + prediction = Clustering + Forecasting [OK]

Hint: Use clustering for segments, forecasting for trends [OK]

Common Mistakes:

• Using highlight tables or filters for segmentation
• Confusing trend lines with forecasting
• Choosing unrelated features like maps or parameters