Six-Sigma is a standard way to score how consistent a process
is. It compares the spread of actual outcomes against an
acceptable range. Higher σ levels = more predictable.
How we apply it here
This dashboard uses the parking lot's vehicle dwell time
as the process variable. Because we don't know your operational
targets, we use the clip's own statistics as the spec:
- Target = median dwell time across all observed vehicles
in this clip.
- Spec band = median ± 2 × IQR (the inter-quartile
range — a robust spread measure).
- Defect = a vehicle whose dwell falls outside that band.
With real operational targets (e.g. "average dwell must be
under 30 minutes"), every number on this panel updates
automatically.
The five metrics
- Sigma level (σ) — overall process quality score.
Higher is better. Industrial benchmarks: 3σ ≈ acceptable,
4σ ≈ good, 6σ ≈ world-class.
- Process capability (Cp) — ratio of spec width to
process spread. Cp ≥ 1.33 is usually considered capable.
Cpk additionally accounts for whether the process is
centred on target.
- Variation (σ) — one standard deviation of dwell
time, in seconds. Lower = more predictable.
- Defect rate — percent of observations that fell
outside the spec band.
- On-time compliance — the inverse — percent that
stayed inside the spec.
Why "illustrative"
Real Six-Sigma analysis requires industry-set spec
limits. We use clip-derived limits so you can see how outliers
(the longest-staying, shortest-staying, or most-erratic
vehicles) drive the score. Replace the spec with your own
targets and the numbers become directly comparable to your
operational benchmarks.