Introduction

At the start of 2023, Sentiance embarked on an exciting new chapter: enhancing motorcycle safety specifically for the gig economy. With our on-device technology, we introduced an affordable and scalable solution tailored to meet the unique needs of this growing sector. Recognizing that 125cc motorcycles are the preferred mode of transportation in the MENA region, where nearly 3,000 companies offered motorcycle delivery services in the UAE alone (a 40% growth from 2021 to 2022), we aimed to address rising safety concerns proactively.

This rapid industry expansion, however, has led to a concerning increase in motorcycle-related incidents. Dubai Police reported that in the UAE, accidents alone grew by nearly 33% from 2020 to 2021. According to Road Safety UAE, 42 deaths and 1,020 injuries in 2023 were attributed to motorcycles. Our mission remains clear: leverage advanced motion insights and AI-driven data science to protect riders while delivering measurable returns for insurers, mobility providers, and safety organizations, ensuring safety initiatives are both impactful and economically viable.

Recognizing the gig-economy opportunity

The gig economy thrives on agile, cost-effective transportation in the MENA region, making 125cc motorcycles incredibly popular. However, the dynamics of gig riding introduce specific risks not typically encountered in other transportation modes. Early on, we identified a pressing need: a solution that not only monitors riding behavior but also adapts to the distinct operational context of gig drivers. With this insight, our team set out to innovate and tailor our technology to meet these challenges head-on.

Detecting two-wheeler trips: The transport classifier upgrade

Accurate detection of two-wheeler trips became the foundation of our new approach. Traditional classifiers were not sufficient for distinguishing the unique patterns of gig economy motorcycle trips. We updated our transport classifier by integrating a “gig flavor”—a set of rules and algorithms fine-tuned to the idiosyncrasies of these journeys. This upgrade allowed us to confidently detect two-wheeler trips, ensuring our system was primed to collect relevant data right from the start.

Engineering for the unique dynamics of motorcycles

Motorcycles present challenges distinct from those of cars, especially when it comes to measuring harsh events. Riding dynamics on a two-wheeler are influenced by factors like countersteering, and phone positioning becomes critical. The picture below shows the effects on magnitude measured by the phone accelerometer when the phone is placed in different locations:

• TEST005 (orange) - right lower arm (android)
• TEST003 (dark green) - left upper arm (iOS)
• TEST006 (blue) - left chest pocket (android)
• TEST008 (light green) - right pants pocket (android)
• The majority of the phones showing the highest magnitudes (in grey) were mounted to the steering wheel

To address these issues, we developed a solution tuned for mounted phones that intelligently filters out misleading turns caused by the natural handling of a motorcycle. We found that nearly all gig economy riders in our client population mount their phones to use them for navigation purposes, thus making this possible. This innovation means our system focuses on genuine risky behaviors—capturing harsh braking and rapid acceleration that truly signal potential danger. 

Beyond the ride: Measuring legal compliance and in-motion communication

Safety isn’t just about handling the ride—it’s also about adhering to legal standards and reducing distractions during transit. Our solution now measures legal driving and tracks call-while-moving behaviors. To track legal driving currently, we look at speeding violations (driving above posted speed limits). Since our solution is on-device we build a Tile Server - a solution that sends a compressed speed limit map to the rider's phone. We also added the ability to correct and override speed limits that we extract from Open Street Maps.

By doing so, we provide a more comprehensive profile of rider behavior, ensuring that our insights support both regulatory compliance and everyday safety. This dual focus reinforces our commitment to proactive safety interventions and legal accountability.

A real-world success story: The client experience

Theory meets reality in the success story of one of our key partners in the gig economy space. Faced with rising accident rates, they turned to our solution and witnessed a tangible reduction in accidents. By leveraging our real-time insights and predictive analytics, they were able to implement various incentives that enhanced rider safety and thus brand image. This transformation serves as a powerful testament to the impact of data-driven safety interventions.

Note that we are not only capable of identifying risky drivers, but also are doing analytics on populations that expose risky regions, cities, areas, and exact locations that require special attention.

Data-driven Crash Detection: From claims to clarity

To further our goal of accident prevention, we began gathering claims data directly from our clients. This collaboration led to an intensive manual labeling effort, where we meticulously analyzed over 700 real crash examples. The resulting dataset empowered us to develop a sophisticated (neural network-based and on-device) Crash Detection solution tuned for identifying and dismissing fraudulent claims—one of the problems gig economy clients face.

Unveiling risk models and fatigue factor

Our journey didn’t stop at Crash Detection. We used the same rich dataset to build predictive risk models that offer deep insights into rider behavior. One key revelation was the critical role of fatigue as a leading risk factor. Riders who ride more than 10 hours a day or/and are working without a break for weeks are 3-4x more likely to be in an accident even if we account for the fact that they have more exposure. Furthermore, our analysis demonstrated that our legal and smooth scoring solutions each provided a lift of up to 2x, reinforcing their effectiveness in differentiating safe from risky riding patterns. These findings not only inform our current practices but also pave the way for future innovations in rider safety.

Conclusion: Driving proactive safety in the gig economy

Our journey—from identifying a niche market need to deploying advanced motion insights and risk modeling—underscores Sentiance’s commitment to rider safety. By blending cutting-edge data science with real-world applications, we have built a solution that transforms how gig-economy motorcycle safety is approached. We invite insurers, mobility providers, and safety advocates to join us in driving the future of rider protection. Let’s connect, collaborate, and create safer roads for everyone.