Leveraging road safety analytics to accelerate Vision Zero goals

How can authorities and road safety practitioners identify potentially unsafe urban locations before accidents actually happen? Relying on historical crash databases alone leaves municipal teams trapped in a reactive loop. To address this challenge, a groundbreaking data-driven approach has been introduced, combining big data and artificial intelligence to shift urban safety from a reactive practice to a proactive strategy. 

This case study outlines how SOTERIA, an innovative road safety analytics platform, transforms complex mobility data into actionable insights that help cities prevent accidents and accelerate Vision Zero objectives.

Overcoming the limits of reactive hotspot analysis

Most conventional hotspot analyses still rely almost exclusively on historical accident databases. While this traditional approach remains useful, it presents clear and critical limitations for city management and long-term Vision Zero planning:

• Reactive approach: analyses based on historical databases only reflect crashes that have already been officially recorded. This means that locations are only flagged as hazardous after an accident has occurred.
• Operational inefficiencies: conventional analyses demand a significant amount of manual data processing effort.
• Static insights: updates are not frequent, making it difficult to maintain continuous situational awareness.

Crucially, the absence of recorded crashes does not necessarily indicate that a road segment is safe. It may simply have avoided the right combination of risk and exposure up to that point. The core challenge for road authorities is discovering how to proactively identify these potentially unsafe locations before accidents occur.

The solution: A data-driven approach to predictive road safety analytics

To overcome legacy limitations, SOTERIA combines traditional accident data with new mobility datasets, shifting the paradigm towards more informed proactive analysis. 

The implementation of the platform revolves around three core technical pillars:

• Connected vehicle data integration: the tool captures information on vehicle position, speed, and acceleration. This makes it possible to detect harsh driving events, such as sudden braking or sharp manoeuvres, which act as vital indicators of risky situations and near-miss conditions.
• Mobile network data utilisation: anonymised mobile network data are used to estimate travel demand. This allows crash patterns to be analysed directly in relation to traffic exposure for various user types.
• AI-powered road crash risk prediction: the platform leverages artificial intelligence to support the prediction of road crash risk. By combining historical accident data, mobility information, near misses, and infrastructure data, the AI estimates precisely where and when accidents are more likely to occur.

A modular software architecture for urban safety

SOTERIA’s software architecture is built around several complementary modules designed to optimize road safety analytics across distinct operational layers:

Descriptive and diagnostic analysis:

• Accidents: visualisation of recorded crashes and their specific attributes.
• Hotspots: identification of high-risk locations grouped by severity and user type.

Exposure and demand analysis:

• Travel demand: characterisation of traffic demand across the urban network.
• Accident exposure: analysing risk relative to vehicle and pedestrian exposure indicators.

Proactive and predictive analysis:

• Dangerous locations: near‑miss detection using Connected Vehicle Data to find high-risk areas.
• Accident prediction: AI-based risk estimation to forecast temporal and spatial accident probabilities.

Delivering measurable road safety outcomes for cities

The SOTERIA prototype was tested in Madrid’s Living Lab alongside municipal police validating its capacity to guide actual urban safety strategies. By merging diverse data streams into a centralized digital environment, the platform effectively minimizes operational silos and reduces the burden of manual data handling.

Transitioning from reactive methods to a predictive framework provides significantl advantages for authorities pursuing Vision Zero objectives:

• Streamlined operations: achieving greater efficiency and speed in the analytical workflows and decision cycles for safety experts.
• Unprecedented visibility: direct access to new urban insights that were previously difficult or impossible to obtain through traditional methods.
• Targeted interventions: better prioritisation of preventive interventions and infrastructure adjustments based on real-time risk assessments, coupled with the ability to evaluate the effectiveness of implemented measures. 

What we learnt from real-world validation

The development and deployment of SOTERIA’s prototype in Madrid provided critical insights into the future of urban road safety analytics:  

• Integration is key: one of the central strengths observed was the ability to integrate multiple, different data sources into a single platform. This consolidation is essential for eliminating fragmented workflows.  
• The power of proxy indicators: relying on near-misses and traffic insights alongside accident data successfully supports the conceptual shift from reactive problem-solving to a proactive understanding of road network safety.  
• Real-world viability: validating the tool in the Madrid Living Lab alongside local police confirmed its operational potential to support genuine, day-to-day decision-making in a major urban environment.  
• Scalability of the methodology: the framework proved that this proactive, data-driven approach is inherently designed to be transferable to other cities and contexts worldwide, providing a scalable blueprint for global Vision Zero strategies.

The SOTERIA’s project

SOTERIA was a European research project aiming to accelerate the achievement of the EU’s Vision Zero goal by improving the safety of vulnerable road users (VRUs). It integrated heterogeneous mobility data, AI, modelling and simulation, and innovative digital services to better understand road safety risks, anticipate hazardous situations, and support safer and more inclusive mobility. The solutions of the project were co-created and tested through four Living Labs across Europe, covering different urban contexts and user groups, including pedestrians, cyclists, motorcyclists, young people and older adults.

The SOTERIA consortium brought together 15 multidisciplinary European partners, combining expertise from research institutions, technology providers, industry specialists, and public authorities.

If you want to learn more about the SOTERIA project, access our R&D project page.