AI Mobility Data: Practical Playbook for Smarter Cities

A modern city can measure traffic in near real time, predict where congestion will pop up next, and test policy options before a single sign is installed. Yet most municipalities still make mobility decisions with lagging indicators: annual counts, infrequent surveys, and complaints that arrive after the damage is done.

Mobility data changes that. And when you combine mobility data with AI, cities get something even more valuable than “visibility”: decision support that’s fast enough to matter—for traffic management, emissions, equity, and day-to-day operations. This post is part of our “Mākslīgais intelekts publiskajā sektorā un viedajās pilsētās” series, where we focus on what actually works in public sector AI, not just what sounds impressive.

The spark for this piece comes from a SmartCitiesWorld podcast conversation with TomTom’s Ralf-Peter Schäfer, who’s spent two decades in mobility and has watched the industry shift from devices to data services and partnerships. The big idea is simple: real-time traffic and mapping data can become a foundation for policy—if cities build the capability to turn data into action.

Why mobility data is now a public sector “must-have”

Answer first: Mobility data is essential because it reveals where the system fails (bottlenecks), who is affected (distributional impacts), and what interventions will work (before-and-after evaluation).

Cities used to treat traffic data as a technical input for signal timing. That’s outdated. The same streams of information—vehicle speeds, travel times, incident patterns, routing behavior—can support public sector goals that sit well beyond transport departments:

• Economic productivity: Less time in traffic improves reliability for workers and logistics.
• Public health: Smoother traffic and mode shift can reduce local air pollution hotspots.
• Climate targets: Emissions are tightly tied to stop-start conditions and network inefficiency.
• Social equity: Travel time burden is often uneven across neighborhoods.

Here’s the stance I’ll take: If a city has climate commitments and growth pressure, mobility data isn’t optional. What’s optional is whether the city is using it responsibly and effectively.

Real-time beats “perfect”

Mobility decisions often stall while teams chase perfect datasets. But for operational questions—“Where are the recurring delays?” “Which corridor is failing every weekday?”—timely, consistent data beats perfect data that arrives too late.

This is where data providers that have evolved into B2B services (as discussed in the podcast) matter. They can offer standardized feeds and historical baselines that municipalities rarely have the budget or staffing to build alone.

From dashboards to decisions: what AI adds to mobility analytics

Answer first: AI turns mobility data into predictions, classifications, and recommended actions—and it scales analysis from single intersections to entire networks.

Plenty of cities already have dashboards. The problem is that dashboards often create a new bottleneck: humans. Analysts can’t manually interpret thousands of road segments across multiple time windows, then translate that into policy options, stakeholder narratives, and budget proposals.

AI helps in three practical ways.

1) Predict congestion and reliability, not just report it

If your data only tells you what happened yesterday, you’re always behind. With machine learning models trained on historical patterns plus real-time signals, cities can:

• Forecast travel time reliability for key corridors (commuter routes, bus corridors, freight access)
• Predict incident impact propagation (how a crash affects upstream/downstream flow)
• Trigger proactive responses (dynamic signal plans, rerouting guidance, operator alerts)

This matters because reliability is what residents feel. A 25-minute trip that sometimes takes 60 is worse than a steady 35.

2) Automate detection of “policy-relevant” patterns

AI classification can identify patterns that are easy to miss:

• Recurrent bottlenecks that only appear during school drop-off windows
• Construction impacts that persist longer than planned
• Seasonal patterns (December retail surges, winter weather disruptions)

And because it’s December 2025: this is the month when many cities see the sharpest conflict between shopping traffic, deliveries, and transit reliability. AI can help operations teams separate “normal seasonal stress” from truly abnormal breakdowns.

3) Simulate options before you commit politically

Policy cycles are short, infrastructure cycles are long, and public attention is unforgiving. AI-supported scenario modeling (paired with good base maps and traffic data) can test:

• Bus lane timing changes
• Low-traffic neighborhood filters
• Dynamic curb rules (delivery windows, ride-hail pickup zones)
• Signal coordination strategies

No model is a crystal ball. But a decent model that’s explained well is often enough to secure the political mandate for pilots.

Snippet-worthy truth: Mobility AI isn’t about “smart roads.” It’s about faster, more defensible decisions in a system where every change has winners and losers.

Equity and environment: using mobility data without fooling yourself

Answer first: Mobility data can support equity and emissions goals only if cities define the outcome metrics up front and track distributional impacts neighborhood by neighborhood.

The podcast highlights how data can help cities analyze travel behavior, promote social equity, and support environmental goals. That’s true—but it’s also where cities get sloppy.

Two common failure modes:

1. Average improvements that hide inequity
   • Citywide congestion drops, but travel times worsen for outer districts.
2. Emissions claims without measurement discipline
   • A project is labeled “green” without tracking stop-start reduction, bus speeds, or mode shift.

A practical equity measurement approach

You don’t need a perfect social model to start. You need consistency and transparency.

1. Pick 3–5 equity-relevant indicators (and publish them)
   • Peak-period travel time to jobs
   • Bus corridor speed and on-time performance
   • Safe walking access to essential services
   • Crash risk proxies (speeding, conflict points)
2. Break results down geographically
   • Districts or neighborhoods, not just citywide averages
3. Commit to “no hidden tradeoffs” reporting
   • If one area gets worse, say it plainly and explain the mitigation plan

Environmental metrics that actually reflect mobility reality

If your goal is emissions reduction through traffic management, watch metrics tied to fuel burn:

• Delay minutes per kilometer on key corridors
• Stop frequency (proxy for stop-start driving)
• Average speed distributions (not just a single average)

And if your goal is mode shift, track:

• Bus travel time competitiveness vs. car
• Reliability improvements after signal priority
• Curb management impacts on cycling and walking comfort

Public-private partnerships: the trust layer cities can’t skip

Answer first: Partnerships work when cities keep authority over outcomes and governance, while vendors provide data, tooling, and operational expertise.

The podcast underscores collaboration and trust: bridging government authorities, tech providers, and the public. That’s not a soft theme—it’s the make-or-break factor.

Here’s what I’ve seen derail smart city mobility programs: a city buys data, but not a decision process. The tool becomes a report generator. The partnership becomes a subscription.

What “good” looks like in a mobility data partnership

A practical checklist cities can use in procurement and program design:

• Clear purpose statement: “We will use mobility data to reduce bus delay by X% on these corridors” (not “improve mobility”).
• Data governance: retention, access controls, audit logs, and public documentation.
• Privacy-by-design: aggregation thresholds, minimization, and strict limits on re-identification risk.
• Interoperability: exports/APIs so insights aren’t locked inside one platform.
• Shared accountability: vendor SLAs for data quality and uptime; city SLAs for operational response.

Building public legitimacy (especially for AI)

Public sector AI fails when it’s treated as a black box. For mobility analytics, cities should communicate:

• What data is used (and what is not used)
• What decisions the AI can influence (and what it can’t)
• How residents can challenge outcomes (appeals, feedback loops)

If a city can’t explain the system simply, it won’t survive the first controversy.

A realistic implementation roadmap for 2026 budgets

Answer first: Start with one operational use case, prove value in 90 days, then expand to planning and policy—while formalizing governance early.

Many municipalities are setting 2026 priorities right now. Here’s a roadmap that respects political cycles and staffing constraints.

Phase 1 (0–90 days): operational wins

Pick one corridor or area and one measurable target.

• Set a baseline: peak travel time, bus speed, incident response time
• Implement monitoring: congestion hotspots, recurring delay windows
• Establish a response playbook: what operators do when thresholds are hit

Deliverable: a short public-facing report that shows before/after change on one outcome metric.

Phase 2 (3–6 months): AI-assisted prioritization

• Add prediction: next-day reliability forecasts for priority corridors
• Automate anomaly detection: construction, events, weather-related disruptions
• Use AI to rank interventions by expected impact and feasibility

Deliverable: a prioritized pipeline of interventions tied to specific metrics and budgets.

Phase 3 (6–12 months): scale + governance maturity

• Expand to more corridors and multimodal metrics
• Integrate with city systems: signals, transit operations, curb management
• Publish a governance framework: privacy, bias checks, procurement rules

Deliverable: an annual mobility outcomes dashboard plus a governance memo that procurement teams can reuse.

People also ask: quick, practical answers

Can mobility data help without building a “smart city platform”?

Yes. Start with a narrowly scoped data feed and a clear operational question. Platforms can come later. The early win is process change, not architecture.

Is AI required to benefit from mobility data?

No. Basic analytics can deliver value. AI becomes useful when scale and speed matter—citywide monitoring, prediction, and scenario testing.

How do we avoid “data-driven theater”?

Tie every dashboard to a decision and an owner. If no one is responsible for acting on an alert, the alert is noise.

Where this fits in the AI public sector story

Mobility is one of the most tangible places where mākslīgais intelekts publiskajā sektorā earns trust: residents feel shorter delays, more reliable public transport, and safer streets. But it only works when cities treat data as part of governance, not just technology.

If you’re shaping a smart city roadmap for 2026, start here: pick one mobility outcome, instrument it with trustworthy data, apply AI where it saves staff time or improves reliability, and publish results in plain language. Then scale.

What would change in your city if every mobility project had to prove—within 90 days—how it improved travel time reliability and who benefited most?