AI-Driven urban key performance indicators (KPIs) - Harnessing non-traditional data and artificial  
intelligence to measure city performance with 
unprecedented accuracy, granularity and timeliness
Foreword
Acknowledgements
Contents
Abbreviations and acronyms
Executive Summary
1 Introduction: Why Cities Must Reinvent Measurement
     1.1 The measurement gap in modern city management
     1.2 A new possibility: The continuously measured city
     1.3 Purpose and structure of this document
2 The Non-Traditional Data Landscape
     2.1 Satellite and aerial earth observation
     2.2 Drone and UAV imagery
     2.3 Mobile network and location data
     2.4 IoT sensor networks
     2.5 Connected vehicles and mobility platforms
     2.6 Street-level imagery and video
     2.7 Transactional and administrative digital exhaust
     2.8 Social media, web and crowdsourced data
3 The AI Toolkit: From Raw Signals to Trusted Indicators
     3.1 Computer vision and deep learning
     3.2 Supervised machine learning for estimation and inference
     3.3 Time-series forecasting and predictive analytics
     3.4 Natural language processing (NLP) and large language models (LLMs)
     3.5 Data fusion and multimodal learning
     3.6 Edge AI and real-time processing
     3.7 Digital twins and simulation
     3.8 Generative and agentic AI
4 From Data to Decision: The AI-Driven KPI Value Chain
5 Where It Works: Application Domains Across the City
     5.1 Mobility and transportation
     5.2 Environment, climate and energy
     5.3 Urban planning, land use and housing
     5.4 Infrastructure and utilities
     5.5 Public safety, resilience and emergency management
     5.6 Economy, tourism and urban vibrancy
     5.7 Health, wellbeing and liveability
     5.8 Governance and service delivery
6 From Concept to Practice: Case Studies
     6.1 Population: Dubai Population Now (United Arab Emirates)
     6.2 Health: dengue forecasting in Ho Chi Minh City and Ha Noi (Viet Nam)
     6.3 Mobility and accessibility: Project Sidewalk
     6.4 What the cases show
7 The Value Proposition: Why This Matters
8 Navigating the Challenges: Risks and Safeguards
     8.1 Privacy and social licence
     8.2 Bias and representativeness
     8.3 Accuracy, validation and model drift
     8.4 Interoperability, legacy systems and vendor lock-in
     8.5 Skills, institutions and regulation
9 A Roadmap for Aspiring Cities
     9.1 A maturity model for AI-driven measurement
     9.2 A phased implementation path
10 Conclusion: From Periodic Reports to a Self‑Aware City