Section 1 — What you will learn
✓ By the end you will understand
- ε (epsilon): the “friend radius” around a point.
- minPts: how many friends are needed to call it a crowd.
- Core point: has enough neighbors inside ε (crowd leader).
- Border point: not a leader, but attached to a leader’s crowd.
- Noise / Outlier: not connected to any crowd.
- Why DBSCAN is special: it finds weird shapes (moons / rings) and ignores outliers.
How to play: Use Step to watch DBSCAN think slowly. Use Auto Mode for a cinematic scan.
Click inside the arena to add a new point live (then press Reset Scan).
Click inside the arena to add a new point live (then press Reset Scan).
Section 2 — Controls
Challenge rewards: correct clustering + low noise mistakes + stable groups.
Story helps kids remember what “noise/outlier” means.
Think: “How far can two points be and still be friends?”
Higher minPts = fewer leaders = more points become border/noise.
Noise is visual: storm lines + jitter + extra rogue points.
Section 3 — Live stats
Clusters Found
—
Noise / Outliers
—
Core Points
—
Border Points
—
Progress
—
Score
0
Live Chart
Assigned vs Noise trend (updates live)
Section 4 — Live Math (simple)
Loading…
🧠 Kid Coach
Loading…
Tip: ε controls friendships. minPts controls “leader strictness”.