When Routing Decisions Increase Risk: A New Way to Keep Buses Safe and Punctual

It’s 7:30 a.m., and the first Route 22 bus pulls out on schedule. Traffic is steady, passengers are calm, and the trip ends with a clean safety report. 

Five hours later, the same driver runs the same route again—but this time, it’s rush hour. Congestion builds up, passengers grow impatient, and a few harsh braking alerts appear in the trip summary. 

For years, operations teams could see individual events like speeding, braking, and cornering—but not the underlying patterns. Was the safety event caused by the driver, the vehicle, the time of day, or the route itself? 

The Missing Piece in Fleet Safety 

Bus and coach executives know how much pressure scheduling decisions create. A few minutes too little in the route plan, and drivers feel compelled to rush. A few too many, and buses crawl between stops, wasting fuel and idling unnecessarily. 

Traditional safety systems could flag risky driving—but they couldn’t reveal whether the route itself or other contextual factors were setting drivers up for failure. 

That’s where GreenRoad’s new AskMila™ “Risk by Route” (RbR) capability comes in. 

Route Intelligence—Driven by AskMila™ 

AskMila™, GreenRoad’s conversational AI assistant for fleet safety leaders, takes data already collected in your GreenRoad system—trip times, safety events, driver performance—and transforms it into route-level insight. 

With Risk by Route, managers can: 

• Compare safety performance across time and conditions (e.g., 7:30 a.m. vs. 12:00 p.m.). 

• Spot high-risk routes where harsh safety events such as braking or speeding events cluster. 

• Identify consistent safety patterns across shifts, vehicles, or depots. 

• Support fairer driver coaching and scheduling decisions based on data, not perception. 

Research confirms what operators already know: environment and timing strongly influence safety outcomes. 

What the Data Shows 

According to Transport for London’s Bus Safety Standard: Bus Braking Data Analysis (2023), London’s buses record roughly 2.4 billion braking events annually, with a notable share of severe decelerations (over 5 m/s²) often linked to congestion and stop density. 

A 2024 study by Ziakopoulos et al. found that road geometry and traffic flow patterns directly affect harsh braking, showing that a small number of routes or segments often account for a large portion of safety events. 

It’s not about punishing drivers—it’s about understanding context.
Two trips on the same route can look entirely different depending on time of day, congestion, and workload. AskMila™ makes these differences visible—so you can act on them. 

For Passenger Transport, Route-Level Risk Insight Changes Everything 

• Targeted safety focus: Identify and fix the few routes responsible for most incidents. 

• Fair driver evaluation: Assess performance under identical route conditions. 

• Smarter scheduling: Align travel times with real-world conditions to reduce pressure and lateness. 

• Lower total cost: Fewer claims, reduced idling and wear and tear—coupled with calmer, safer journeys. 

Once operators adjust schedules to reflect reality, both safety alerts and stress levels drop.
It’s not about slowing down—it’s about matching routing decisions to reality. 

Smarter Decisions, Simple Questions 

AskMila™ makes complex analysis effortless. There’s no need to run reports—just ask: 

• “Show me the average safety score for Route 301 from Thursday 21 to Thursday 28.” 

• “Which route has the highest number of safety events?” 

• “What are the three best- and worst-performing routes in ?” 

Each answer helps safety and operations leaders see how time, traffic, and workload interact—enabling smarter, more realistic planning for the next shift. 

The data is clear: understanding route-level risk leads to safer, smoother, and more efficient operations. 

References 

1. Transport for London, Bus Safety Standard: Bus Braking Data Analysis, 2023. Available here (PDF) 

1. Ziakopoulos, A., et al. (2024). Analysis of harsh braking and harsh acceleration occurrence via explainable imbalanced machine learning using high-resolution smartphone telematics and traffic data. ResearchGate