ClusteringPRO — Exploring Athlete Data with Structure and Meaning

If you’re working in sport — especially as a strength coach, performance analyst, or sports scientist — you’ve likely found yourself staring at spreadsheets full of numbers, wondering:

“What do I do with this?”
“Which variables matter?”
“Are these athletes really that different?”

Data is everywhere now — the challenge isn’t collecting it, it’s making sense of it. We don’t need dashboards filled with red-yellow-green; we need structure, insight, and direction. That’s exactly why I built ClusteringPRO — a tool to help you explore multivariate data, group athletes meaningfully, reduce complexity, and find patterns that are actually interpretable.

🎯 What ClusteringPRO Does

ClusteringPRO is a Shiny app for Complementary Training members that provides an exploratory data workflow built around:

✅ Hierarchical clustering of athletes
✅ PCA (Principal Component Analysis) for dimensionality reduction and visual interpretation
✅ Variable clustering using multiple methods
✅ Archetypal analysis
✅ Custom variable scaling and weights — emphasize what matters in your context
✅ Visualizations that are actually helpful: dendrograms, heatmaps, PCA scatter plots, and archetype loadings

You can run the full pipeline in a few clicks — no coding, no prep scripts.

🧪 Built-In Example Datasets — Start Exploring Immediately

To make it easier for coaches and analysts to understand these methods without fear of breaking something, ClusteringPRO includes multiple built-in demo datasets.

These were created with purpose:

Simple datasets with no clusters — to show what real randomness looks like
Structured datasets with obvious groupings — to help interpret clustering outcomes
Archetypal geometries (e.g., triangle, pyramid) — to visualize what extreme profiles “look” like
Real-world inspired samples for hands-on learning

You can toggle through them, apply the full analysis workflow, and see how the algorithms respond.

This makes ClusteringPRO not just an analysis tool, but a learning platform.

🧱 Clusters vs. Archetypes — What’s the Difference?

Let’s clear something up — clustering and archetypal analysis often get lumped together, but they serve different purposes.

Concept	What It Does	How You Might Use It in Sport
Clustering	Groups similar observations together based on overall similarity	Segment athletes into types (e.g., “power-speed group”)
Archetypes	Finds “extreme” or pure profiles on the edges of your data	Identify idealized profiles — “power prototype,” “endurance prototype,” etc.
Interpretation	Clusters are internal structure; archetypes are external boundaries	Use clusters to assign, archetypes to inspire or compare
Blending	Each athlete belongs to one cluster; but may be a blend of archetypes	An athlete might be 70% archetype A, 30% B

You can think of it like this: clustering answers “who belongs together?”, while archetypes answer “what are the pure types that exist?”

You can use clusters to guide team segmentation or training prescription — and archetypes to highlight extremes, role models, or deviations from “ideal.”

🏋️ Who This Is For

ClusteringPRO is built for real-world practitioners — not data scientists in a lab:

🏋️‍♂️ S&C Coaches — Who want to group athletes or explore outliers
🧪 Sports Scientists — Looking to simplify or restructure large test batteries
📊 Performance Analysts — Seeking insights into multivariate profiles
📚 Researchers or PhD students — Doing exploratory work on team data

You don’t need to touch a single line of R code. You just need a .csv, and a question.

📈 Example Use Cases

Profiling Camp: Cluster athletes post-test into interpretable groups, then compare against archetypes for insight
Monitoring Review: Cluster variables to identify redundancies (e.g., does HRV say anything new?)
Report Visualization: Use PCA and dendrograms to add structure to what’s normally a messy testing sheet
Test Reduction: Use variable clustering to identify which metrics overlap — and which truly contribute

🎥 Instructional Videos — From Overview to Deep Dive

To help you get started, I’ve recorded two videos:

Quick Overview — A walkthrough to give you a feel for what the app can do
Full Tutorial (in the making) — A deeper instructional video showing how to:
- Upload and preprocess data
- Apply clustering and archetypal analysis
- Interpret results
- Use variable weighting and custom scaling