Fleet Fuel Cost and Idle Time Analysis Builder

You are a data analyst helping an operations team reduce fleet fuel cost and idle time responsibly. Analyze the available data for:

Fleet type: [delivery vans, service vehicles, trucks, buses, mixed fleet]
Business context: [last-mile delivery, field service, public works, school transport, construction]
Data sources: [fuel card, telematics, GPS, odometer, maintenance, route plan, dispatch, weather]
Time period: [weeks, months, season]
Vehicle fields: [vehicle ID, type, age, fuel type, MPG, engine hours, maintenance status]
Driver fields: [driver ID or anonymized group, route, shift, training status]
Trip fields: [miles, stops, idle minutes, harsh events, speed, route duration, payload if available]
Fuel fields: [gallons, cost, station, date, odometer, transaction exceptions]
Operational constraints: [customer time windows, safety policy, union rules, PTO, geography, equipment needs]
Known concerns: [high idle time, fuel theft, route inefficiency, maintenance issues, winter warmups]
Decision needed: [coaching, route redesign, maintenance, policy update, replacement planning]
Privacy and fairness needs: [driver anonymization, coaching boundaries, weather/route normalization]

Create:
1. Data quality checks for fuel card, odometer, telematics, and route data.
2. KPI definitions for MPG, cost per mile, idle minutes per engine hour, fuel exceptions, and utilization.
3. Segmentation plan by vehicle type, route, driver group, shift, geography, and season.
4. Outlier analysis for high fuel cost, high idle time, impossible odometer readings, and suspicious transactions.
5. Normalization guidance for weather, payload, terrain, vehicle age, route density, and customer requirements.
6. Root-cause hypotheses for fuel waste, maintenance issues, dispatch patterns, training gaps, and policy constraints.
7. Savings opportunity table with confidence level, operational owner, expected impact, and risk.
8. Dashboard layout for executives, fleet managers, dispatch, and maintenance teams.
9. Driver coaching guardrails that avoid unsafe incentives or unfair comparisons.
10. Recommended next experiments and monitoring plan.

Do not recommend actions that compromise safety or ignore route constraints. Mark missing data and fairness caveats clearly.