HCT · INTERNAL MANAGEMENT
Diesel integrity trends & SPC
Loading selected dataset…
Monthly fleet comparison
Negative: GPS rise is lower than purchased litres. Positive: GPS rise is higher. Neither sign is a pass/fail result. Zero is exact agreement; historical control limits describe variation. Management tolerance has not been defined.
Each session, including every member purchase, is assigned to its existing Malaysia-time session month. GPS rise is counted once. Coverage is included purchase volume / all purchase volume on that same basis; it does not establish uninterrupted GPS coverage or accuracy.
Lorry-by-month heatmap
Selected-month comparison
Fixed fleet reference — July–September 2026
Provisional fleet reference limits are investigation thresholds. An observation inside the range is not proof of accuracy. Values remain fixed until an explicitly calculated replacement is reviewed and approved.
Session variation by lorry
Purple outline / hollow square: excluded from baseline fitting; still evaluated against limits. Red points: outside approved fleet reference range. Blue points: box-plot outliers or few-session observations. Fleet lines apply to session percentages only.
Boxes show median and middle 50%; whiskers extend to observations within 1.5 IQR. Outlying points are descriptive, not SPC alarms. Fewer than five sessions: individual points only; this is a display convention.
Contributing sessions and evidence
Excluded sessions and quality reasons
Quantity charts show one lorry at a time within the selected period. SPC uses the existing provisional eligibility rules; fleet measurement filters do not change this baseline. Incident charts show that lorry and the fleet. Each quantity observation represents one matched session. Combined purchase litres and the GPS rise are counted once per session. Limits describe historical variation, not an acceptable error threshold.
Per-lorry quantity difference
Blue: baseline points · teal: eligible monitoring readings · amber diamonds: additional quality warnings · red: provisional anomaly. Click a point to inspect the original purchase records in its session.
Review baseline proposal
Quantity observations and signals
Purchases without detected refill — monthly trend
Red incidents have usable GPS but no detected refill. Missing GPS, unreliable signals and ambiguous matches remain separate. Rate = no-refill purchases / (timing matched + no-refill purchases). Three monthly totals are descriptive; no monthly control limits are fitted.