Designing safer roads in cities through connected vehicle data
A practical guide for local and regional governments
About these guidelines
These guidelines are a practical resource for local and regional governments (LRGs) who want to understand what connected vehicle data is, why it matters for LRGs, how cities are using it today, what stands in the way, and what concrete steps decision-makers can take to get started.
It was developed through a four-session expert working group convened between February and May 2026 by UNDP’s SDG Local Action programme, bringing together practitioners from Buenos Aires, Istanbul, Utrecht, London, Florence, and the Netherlands.
This is not a technology manual. It is a governance and leadership resource for local government officials, transport planners, and elected representatives who must decide whether and how to put this emerging data to work in service of safer, more efficient, and equitable urban mobility.
How to use these guidelines
The guide is organized into six sections that can be read in sequence or consulted individually:
- Section 1 introduces the concept of connected vehicle data and why it is relevant for LRGs.
- Section 2 maps the key use cases cities are already deploying.
- Section 3 summarizes the benefits reported by working group participants.
- Section 4 identifies the main challenges and how practitioners are addressing them.
- Section 5 presents the data-sharing ecosystem: who the key actors are and what roles they play.
- Section 6 offers policy recommendations for cities at different stages of readiness.
Section 1
Why connected vehicle data matters for local governments
The opportunity
We are now starting to see new vehicles — car, bus, bicycle, or scooter — are increasingly equipped with many sensors situated across the vehicles that generate data as whether it is still or if it is moving. A connected vehicle continuously logs its position, speed, acceleration and braking patterns, wheel behaviour, ambient temperature, wiper and light activation, and in some cases air quality readings. Individually, these signals are operational data points. Through improved connectivity and data processing, aggregated across thousands of vehicles moving through a city, they become something far more powerful: a continuous, city-wide sensing layer.
In particular, when it comes to road safety, cities are basing interventions on previous incidents e.g. fatalities or serious injuries but this doesn’t necessarily identify where the incident may have started e.g. it could have started at a different location to where the collision occurred.
VEHICLE DATA matters for three reasons
First, it fills the gaps in existing knowledge. Traditional data collection — fixed cameras, periodic inspections, citizen complaints — is limited in coverage, frequency, and spatial granularity.
Second, it enables a shift from reactive to proactive management. Instead of waiting for a road to fail, an accident to happen, or a citizen to file a complaint, cities can detect deteriorating conditions, emerging risk patterns, and near-miss events in near real time — and act before harm occurs.
Third, it supports better investment decisions. Infrastructure spending is often based on engineering judgement and historical incident data. Connected vehicle data provides objective, evidence-based inputs that help cities direct resources to where they are most needed.
What is a “connected vehicle”?
A connected vehicle is any vehicle that generates and transmits data about its own operation and environment. This includes:
- Passenger cars and light commercial vehicles equipped with telematics or OEM (original equipment manufacturer) connectivity systems;
- Public transport fleets with GPS and onboard sensors;
- Micromobility vehicles; motorized scooters and mopeds; these categories have distinct data and regulatory profiles and should be treated separately where possible;
- Freight and delivery vehicles with telematics systems.
Section 2
Use cases
1. Road safety
Connected vehicle data is used to identify harsh-braking hotspots, understand movement patterns at dangerous junctions, and build predictive road safety risk models that go beyond historical collision data.
Transport for London (TfL) uses aggregated data from fleet operators and OEMs to identify near-miss locations. This data now forms part of TfL’s Vision Zero safety model and is used to measure whether safety interventions are working — faster and more cost-effectively than traditional survey methods.
Istanbul’s Traffic Management Center fuses vehicle CAN bus data with camera feeds to detect incidents in near real time, including accidents at unmonitored locations.
2. Asset management and road conditions
Vehicles act as rolling inspection units, continuously reporting on road surface quality, pavement defects, signage visibility, and road marking legibility. This shifts asset management from periodic, manual inspection to continuous, automated monitoring.
The Dutch ROMO2 programme uses vehicle-generated data to detect pavement deterioration, identify defective traffic signals, and monitor road marking quality — allowing road operators to intervene earlier and prioritize maintenance budgets more accurately.
Istanbul’s Traffic Management Center deployed a surface-scanning device on public buses; the volume of alerts generated revealed the scale of road quality challenges and highlighted the need for intelligent prioritization filters.
Vehicle data can provide a continuous picture of road surface condition, helping cities maintain appropriate standards for all user types — from heavy freight to cyclists.
3. Winter service management
Vehicle data including ambient temperature, wheel-slip signals, and wiper activation allows road operators to identify where and when roads are most slippery — enabling targeted gritting rather than blanket treatment of the entire network.
The Netherlands cities and provinces use this approach to optimize winter maintenance, reducing costs while improving safety outcomes.
4. Micromobility and active mobility
Connected bicycle and micromobility sensors provide data on cycling routes, speeds, air quality exposure, and time spent waiting at traffic lights. This data supports infrastructure investment decisions, traffic light optimization, and behaviour-change programmes. Crucially, investments in cycling infrastructure should be evaluated from a whole-network perspective: the benefits of cycle lane provision frequently accrue to all road users — reducing conflicts, improving transit reliability, and easing car network congestion — not only to cyclists. Data analysis should reflect this systemic perspective rather than treating cycling in isolation.
The Province of Utrecht’s Snuffelbike (Snifferbike) project fitted citizen cyclists with environmental sensors to map air quality and cycling behaviour, enabling evidence-based adjustments to signal timing and route planning.
See.Sense deploys cycling intelligence sensors to measure network performance and usage, helping cities evaluate the return on cycle lane investment and identify where infrastructure gaps are causing unsafe behaviour.
A recurring challenge across both Istanbul and Florence is that cycling infrastructure is often fragmented — well-designed segments interrupted by unresolved crossroads or missing stretches that push cyclists onto carriageways. Connected vehicle data can help identify exactly where these network gaps occur, by detecting changes in cyclist behaviour (speed, trajectory, hesitation) at transition points.
5. Public transit monitoring
GPS and onboard sensors in bus fleets provide real-time vehicle location data, enabling dynamic passenger information systems and improving service reliability.
Real-time bus location data for passenger information is among the most widely deployed applications of connected vehicle data globally, with cities across all income levels and regions having implemented it over many years. Istanbul and Buenos Aires are examples within this working group: both use real-time fleet position data to deliver accurate arrival predictions through passenger-facing applications, supporting modal shift to public transport.
6. Integrated and multimodal transport
The most sophisticated application of connected vehicle data treats the entire transport network — cars, buses, bicycles, pedestrians — as a system to be optimized together with other data sources from the city. Data from one mode informs decisions about others, and investment in cycling infrastructure is evaluated for its impact on car and bus network performance.
The Province of Utrecht emphasizes the importance of a “complete network approach”: the benefits of cycling infrastructure often accrue to all road users, not just cyclists. Data must reflect this systemic perspective.
7. Environmental monitoring
Vehicles equipped with air quality sensors can act as mobile monitoring stations, providing fine-grained spatial data on pollution exposure. This data supports both infrastructure decisions (e.g., routing cyclists away from high-exposure corridors) and behaviour-change programmes.
Section 3
Benefits
Safety improvements
Moving from reactive to predictive road safety management. Multiple sensors allow cities to intervene before collisions occur.
Safety improvements
Road operators can detect deteriorating conditions earlier and prioritize resources more accurately.
New geographic coverage
Gaining visibility into locations that were previously unmonitored, generating hyperlocal data across the entire network.
Behaviour-change tools
Using data to support measurable changes in how citizens travel. It can also support any societal outcome a city prioritises.
Efficient inter-municipal collaboration
Creating shared data infrastructure that allows municipalities to pool resources, share insights, and develop common standards.
Budget savings
Reducing the frequency and cost of manual inspections.
Workforce support
Supporting staff to work more strategically by freeing them from routine data-collection tasks.
Systematic approach
Consolidating previously fragmented point-by-point sensor data into a city-wide picture.
Future readiness
Laying the data infrastructure layer needed for smarter traffic systems, cooperative mobility services, and — in the longer term — autonomous vehicle integration.
Section 4
Challenges
Implementation costs
Accessing, processing, and acting on connected vehicle data requires investment in platforms, technical capacity, and procurement processes.
Mitigation
- Start small with one use case and one data provider.
- Regional pooling (multiple municipalities sharing a single data contract and analytical capacity).
Fragmented governance and data management
Transport data in most cities sits across multiple agencies, each with different systems, mandates, and cultures.
Mitigation
- Identify a lead authority (often at national or regional level) to act as the data ecosystem orchestrator.
Data volume and prioritization
The volume of data generated by even a modest fleet can quickly exceed a city’s analytical capacity.
Mitigation
- Define priority use cases before collecting data.
- Work with data suppliers to apply threshold rules and clustering algorithms that surface only actionable alerts.
Sample representativeness and bias
Connected vehicle data reflects the vehicles that generate it — which may not represent the full road-using population.
Mitigation
- Cross-reference vehicle data counts with independent traffic counts to estimate market penetration by area.
- Ask data suppliers explicitly what vehicle types and geographies are represented in their dataset.
Fragmented partner coordination
Aligning actors in connected vehicle data — particularly in a public procurement context — is complex and time-consuming.
Mitigation
- Structure procurement as an innovation partnership rather than a single contract.
Standardization
Each vehicle manufacturer defines signals slightly differently. Without standardization, combining data from multiple sources is unreliable. Existing standardization processes are often slow and subject to patent-pool dynamics that can create access barriers.
Mitigation
- Adopt existing open standards where they exist.
- Where new standards are needed, build a “coalition of the willing” to develop them openly.
- Advocate for open, royalty-free standards in any multilateral dialogue on connected vehicle data.
Technical limitations
False positives, data latency, connectivity gaps, and uneven technology readiness across use cases are recurring operational challenges.
Mitigation
- Combine multiple data sources to create redundancy.
- Define service levels with data suppliers.
- Accept that some use cases are mature enough for operational use while others require further piloting.
Legal and privacy constraints
Vehicle data can include information that, if not properly anonymized, could be used to identify individuals — raising GDPR and equivalent legal concerns in many jurisdictions.
Mitigation
- Most road management use cases require only aggregated, anonymized data.
- Work with data suppliers to ensure personal data is processed at source. Include data protection obligations in all contracts and define data retention policies from the outset.
Building trust in new data and use cases
Uncertainty about how data will be used, who benefits, and what privacy safeguards are in place can generate resistance from staff, unions, and citizens alike.
Mitigation
- Be transparent from the outset about the purpose of data collection and how findings will be used.
- Demonstrate shared value — make clear what citizens, staff, and partner agencies stand to gain.
- Where unionised workforces may perceive data-driven efficiency as a threat to jobs, frame the programme in terms of workforce redeployment and skill development rather than headcount reduction.
- For new or experimental use cases, pilot openly, share results, and build confidence incrementally before scaling.
Section 5
The data-sharing ecosystem
Key roles in the ecosystem
No single actor holds all the data, capabilities, or authority needed to make it work. Understanding who the different actors are, what roles they play, and how their interests align or conflict is essential for any city approaching this space.
Role
Description
Typical actor
Data generator
Produces raw vehicle data through normal vehicle operation.
Individual drivers and vehicle owners; fleet operators; OEM connected vehicles.
Data owner / holder
Holds legal rights to or custody of the raw data. Often differs from the generator.
Vehicle manufacturers (OEMs); fleet operators; data aggregators; individual vehicle owners.
Data aggregator
Collects data from multiple sources (OEMs, fleet operators), normalizes it, and makes it available to buyers in a standardized format.
Private data platforms; national data access points (e.g., NDW in the Netherlands).
Data processor / analyst
Transforms raw or aggregated data into actionable insights, risk models, or operational alerts.
Technical consultancies; universities; local and regional government’ data teams; AI platform providers (e.g., Snap4City).
Data user / decision-maker
Uses insights derived from the data to make operational, planning, or policy decisions.
City transport departments; road maintenance teams; traffic management centers; elected officials.
Ecosystem orchestrator / coordinator
Defines the governance framework, sets standards, ensures fair access, and aligns the interests of the other actors.
National or regional governments; international bodies (EU); specialist agencies (ROMO programme).
Key actors and their interests
Actor
Role
Interest
Key barrier
Vehicle manufacturers (OEMs)
Role
Data owner / holder
Interest
Monetizing vehicle data; maintaining customer relationships; regulatory compliance
Key barrier
Reluctance to share data openly; proprietary format
Fleet operators
Role
Data generator; data owner
Interest
Operational efficiency; insurance savings; regulatory compliance
Key barrier
Data fragmentation across vehicles and providers
Data aggregators / platforms
Role
Data aggregator; data processor.
Interest
Commercial data brokerage; platform growth; market dominance.
Key barrier
Multi-OEM normalization complexity; accountability gaps.
City transport authorities
Role
Data user; data generator (infrastructure sensors).
Interest
Safer roads; efficient service delivery; evidence-based policy.
Key barrier
Limited technical capacity; budget constraints; siloed governance.
National / regional governments
Role
Ecosystem orchestrator; data user; standard-setter.
Interest
Network-wide safety and efficiency; EU compliance; fiscal responsibility.
Key barrier
Slow procurement processes; limited regional focus.
Citizens / individual vehicle owners
Role
Data generator; data owner (the individual who produces vehicle data holds rights over it — a principle recognised under the EU Data Act and equivalent frameworks); end user.
Interest
Safer roads; cleaner air; better services; privacy protection.
Key barrier
Low awareness; privacy concerns; limited agency in data transactions.
Knowledge institutions / universities
Role
Data processor; researcher.
Interest
Scientific validation; knowledge generation; applied research.
Key barrier
Theory-practice gap; slow technology transfer.
International organizations
Role
Ecosystem orchestrator; standard-setter; capacity builder.
Interest
Global standards; equitable access; SDG alignment.
Key barrier
Implementation distance; coordination complexity.
Section 6
Policy recommendations
Stage 1: getting started
1. Define the problem before acquiring the data
The most common mistake cities make is treating connected vehicle data as a solution in search of a problem. Start with a specific operational or policy challenge and work backward to identify whether vehicle data can help address it.
2. Map your existing culture data
Start by understanding how cultural data is managed in the city. From this, the city can determine the best way to ingest a new data format. In addition, audit what already exists.
3. Secure political and organizational commitment
Data-driven transport management requires buy-in from decision-makers and operational staff. Identify a senior champion.
Stage 2: building capability
4. Build analytical capacity — or borrow it
Cities with limited in-house capacity should partner with universities, regional data teams, or technical consultancies. Regional pooling — as demonstrated by the Province of Utrecht — is one of the most cost-effective approaches for medium and small LRGs.
5. Establish a data governance framework early
Define data ownership, retention policies, privacy obligations, and decision-making responsibilities before signing contracts. Determine who has the authority to act on data-derived insights and who is accountable for the decisions that result. Involve legal counsel in reviewing data agreements with providers.
6. Prepare a data sharing agreement. Insist on open standards and interoperability
Before signing any contract with a data provider, cities should have a clear data sharing agreement in place that defines the terms of access, use, and accountability. Avoid proprietary formats and vendor lock-in. Require data suppliers to provide data in open, documented formats. Contribute to standardization processes. The long-term viability of a connected vehicle data programme depends on the ability to switch providers, add new data sources, and integrate with other urban data platforms.
Stage 3: scaling impact
7. Establish peer-learning and mentoring relationships
Partner organisations should explore creating a structured mentoring mechanism: pairing cities with established programmes with cities in earlier stages, particularly in the Global South. A regional pooling model, where a group of municipalities shares a single data contract and a jointly funded analytical team.
8. Engage the private sector as a partner, not just a vendor
The commercial incentives of OEMs, aggregators, and data platforms need to be shaped by clear regulatory frameworks and strong public sector demand. Cities that organize collectively (at local, regional, national level) have far more leverage in negotiations with industry than individual authorities acting alone.
9. Advocate for global equity in access
Three actions can help cities in the Global South to leverage connected vehicle data.
- Advocate for global OEM data-sharing obligations
- Explore international funding mechanisms to finance consortia of cities in developing regions piloting connected vehicle data applications.
- Ensure that global standardization work is conducted openly and without proprietary lock-in, so that the standards that emerge are accessible to low-resource cities without licensing barriers.
The SDG connection
Connected vehicle data is directly relevant to multiple Sustainable Development Goals.
Contributors
These guidelines are the output of a four-session expert working group convened between February and May 2026.
Participants represented a deliberate mix of city sizes, geographies, and institutional contexts. They are:
- Rikesh Shah, Innovate Today, UK
- Dan Knight, Transport for London, UK
- Erik Vrijens, Dutch Ministry of Infrastructure and Water Management
- Eric van Dijk, Province of Utrecht, Netherlands
- Mustafa Samet, Istanbul Traffic Management Centre, Türkiye
- Tomas Daels, Government of the City of Buenos Aires, Argentina
- Charlie Wilson, See.Sense, UK
- Paolo Nesi, University of Florence/Snap4City, Italy