Example: Clustering Flower Species Data

Business data rarely arrives with clear labels. Customers, products, and markets don’t naturally organize themselves into neat buckets. This is where clustering becomes a powerful analytical technique—helping organizations uncover hidden patterns, segment entities intelligently, and make data-backed decisions at scale.
In this article, we explore clustering and segmentation through practical examples and demonstrate how Tableau’s built-in clustering capabilities make advanced analytics accessible to business users—not just data scientists.

What Is Clustering?
Clustering is the process of grouping similar observations or data points based on shared characteristics. The goal is simple:
Data points within a cluster should be more similar to each other than to those in other clusters.
A Simple Business Example
Consider a car manufacturer analyzing customer preferences:
Cluster 1: Buyers looking for compact cars priced under $6,000
Cluster 2: Buyers seeking spacious cars priced above $30,000
Each cluster represents a distinct market segment with different expectations, pricing sensitivity, and feature requirements. Identifying these clusters enables manufacturers to design targeted models, optimize pricing strategies, and forecast demand more accurately.
This same logic applies across industries—from retail and finance to healthcare and media.

Why Clustering Matters in Business
Clustering helps organizations move from intuition-driven decisions to evidence-based segmentation. It supports:
Customer segmentation for targeted marketing
Product portfolio optimization
Geographic or demographic analysis
Risk profiling and anomaly detection
Demand forecasting and pricing strategies
Instead of asking, “Who do we think our best customers are?”, clustering allows businesses to ask,
“What does the data tell us about naturally occurring groups?”

Clustering in Tableau: Under the Hood
Tableau provides clustering using the K-means algorithm, a centroid-based approach widely used in analytics.
How K-Means Works
Data is divided into K clusters
Each cluster has a centroid (mean position of points)
The algorithm iteratively assigns points to the nearest centroid
The objective is to minimize the total distance between points and their assigned centroid
In simple terms, Tableau automatically groups data points so that similarities within clusters are maximized and differences across clusters are minimized.
The power of Tableau lies in how easily this advanced technique can be applied—often with simple drag-and-drop actions.

Hands-On Example: Clustering Flower Species Data
Let’s walk through a practical example using a flower dataset that includes measurements such as petal length and petal width across three species.
Step 1: Load the Dataset
After loading the dataset into Tableau, review the available features. You’ll notice measurements that can help distinguish between flower species.
Step 2: Create a Scatter Plot
Drag Petal Length to Columns and Petal Width to Rows.
Initially, Tableau aggregates measures by default, resulting in a single data point. To view individual observations:
Go to Analysis
Uncheck Aggregate Measures
You’ll now see a scatter plot representing individual flowers.

Applying Clustering in Tableau
To create clusters:
Open the Analytics pane
Drag Cluster onto the visualization
Tableau automatically creates clusters based on the fields in the view. You can:
Adjust the number of clusters
Choose which variables should influence clustering
This flexibility allows analysts to experiment and validate hypotheses interactively.

Understanding Cluster Quality: Model Interpretation
Creating clusters is only half the job. Understanding why clusters exist is equally important.
Tableau provides a “Describe Clusters” option that explains:
Variables used in the model
Statistical measures validating the clusters
Two critical metrics here are F-statistic and P-value.
F-Statistic (F-Ratio)
The F-statistic measures how well a variable distinguishes between clusters.
F=Between-group variabilityWithin-group variabilityF = \frac{\text{Between-group variability}}{\text{Within-group variability}}F=Within-group variabilityBetween-group variability​
Higher F-values indicate stronger discriminatory power
Variables with higher F-statistics are more influential in forming clusters
P-Value
The P-value indicates statistical significance.
A lower P-value means stronger evidence that cluster means differ
When the P-value falls below a significance threshold, differences are unlikely due to chance
Together, these metrics help validate whether your clusters are meaningful or arbitrary.

Saving Clusters for Further Analysis
Tableau allows clusters to be saved as a dimension:
Drag the Cluster field from Marks to Dimensions
Use it across worksheets, filters, or dashboards
This enables deeper analysis such as performance comparison, trend tracking, and reporting by cluster.
Fields Not Supported in Tableau Clustering
Tableau does not allow the following fields in clustering:
Dates
Bins
Sets
Table calculations
Blended calculations
Ad-hoc calculations
Parameters
Generated latitude/longitude values
Understanding these limitations helps avoid modeling errors early.

Example 2: Country Segmentation Using World Indicators
Tableau’s sample World Indicators dataset provides an excellent real-world use case.
By clustering countries based on indicators such as:
Life expectancy
Urban population
Population over age 65
You can uncover meaningful global patterns.
Insights from Clustering
The U.S. may fall into a high-development cluster
Emerging economies like India or Brazil may cluster together
Regions in Africa or South America often form distinct segments
Selecting a cluster and switching to a text table reveals the countries within it—enabling policy analysis, investment strategy, or market entry planning.

Segmentation Beyond Customers: A Strategic Perspective
While segmentation is commonly associated with customers, it applies equally to:
Products
Markets
Regions
Suppliers
Behavioral patterns
The key is clarity of objective.

A Deeper Segmentation Example: Publishing Industry Case
Imagine a publishing company specializing in business books planning to expand into:
Philosophy
Marketing
Fiction
Biography
Their objective is to identify which age groups prefer which genres.
Step 1: Define the Objective
Clear objectives prevent endless slicing and dicing.
Here, the goal is to identify target age groups for each genre.
Step 2: Identify the Right Data
The dataset includes:
Customer age
Preference scores for each genre
In real-world scenarios, this could be enriched with demographics, geography, or purchase behavior.
Step 3: Create Segments and Micro-Segments
Start with aggregate analysis, then progressively add detail.
Fiction may appear most popular overall
But age-based analysis reveals:
Under 20 → Fiction
20–30 → Business & Marketing
40+ → Philosophy & Biography
Without segmentation, aggregate averages would mask these insights.

Refining the Strategy: Overlapping Preferences
If the publisher can launch only one additional genre, clustering and relationship analysis help refine the choice.
Analysis shows:
Strong overlap between Business and Marketing in the 20–30 age group
Weaker overlap between Business and Philosophy
Final Recommendation
Launch Marketing books targeting customers aged 20–30 years.
This conclusion is driven entirely by data—not assumptions.

Conclusion: From Patterns to Decisions
Clustering transforms raw data into structured insight. But its real value lies not in the algorithm—it lies in interpretation and application.
Key takeaways:
Always start with a clear objective
Validate clusters using statistical indicators
Combine clustering with domain knowledge
Reiterate and refine as new data becomes available
As data grows more complex, tools like Tableau enable business teams to perform sophisticated analysis without heavy statistical expertise.
At Perceptive Analytics, our mission is “to enable businesses to unlock value in data.” For over 20 years, we’ve partnered with more than 100 clients—from Fortune 500 companies to mid-sized firms—to solve complex data analytics challenges. Our services include working with experienced Snowflake Consultants and delivering enterprise-grade Microsoft Power BI consulting services, turning data into strategic insight. We would love to talk to you. Do reach out to us.