Urban Informatics

Urban informatics harnesses the convergence of advanced sensing technologies, ubiquitous data streams, computational methods and geographic information systems to analyse and model the structure, dynamics and services of cities. By integrating data from satellites, fixed and mobile sensors, crowdsourced information and administrative sources, it enables fine-grained monitoring of mobility patterns, infrastructure performance, environmental quality and social interactions. Machine learning and deep learning techniques extract features and predict urban events in real time, while digital twins and virtual models support scenario testing and decision-making in planning and operations. Applications span traffic management and smart parking to pedestrian accessibility, street-level analytics and asset maintenance, all aimed at promoting efficiency, resilience and liveability in rapidly evolving urban environments.

Research from Nature Portfolio

Novel deep-learning frameworks have been developed for real-time detection of critical road infrastructure elements. A stereo depth-camera system combined with an optimised convolutional network achieves over 96 per cent accuracy in identifying manhole covers, offering an embedded solution for vehicle-based road-surface monitoring and enhancing maintenance regimes. Advanced allocation algorithms have been proposed to balance parking demand equitably across urban zones. A suite of heuristic methods, validated over thousands of simulated scenarios, reduces disparities in occupancy and can be executed in milliseconds, guiding dynamic driver allocation and improving utilisation. In response to public-health and distancing imperatives, a network-based analysis of pedestrian pathways across multiple cities demonstrated that shared-effort heuristics can repurpose road space to expand sidewalks without degrading vehicular connectivity, supporting safer, more inclusive public realms during emergencies and beyond.

Urban Informatics publication trend

The graph below shows the total number of articles in urban informatics across all publications each year (not limited to Nature Index journals).

Technical terms

Urban informatics: Interdisciplinary study of cities using data acquisition, analysis and computational methods to understand and optimise urban systems.

Big data: Extremely large and complex datasets generated continuously by sensors, devices and transactions, requiring specialised methods for storage, processing and analysis.

Machine learning: A suite of algorithms that learn predictive or descriptive patterns from data without explicit programming, widely used in urban analytics.

Deep learning: A branch of machine learning employing multilayer neural networks to automatically learn hierarchical features from raw data, often applied to image and signal processing.

Sensor network: A distributed system of interconnected devices that collect, transmit and sometimes process data on environmental or infrastructural conditions in real time.

Spatio-temporal analysis: Techniques that examine how phenomena evolve across space and time, essential for understanding mobility flows, land-use change and environmental dynamics.

References

1. Envisioning the New Urban Informatics.
2. Real-time detection of road manhole covers with a deep learning model. Scientific Reports (2023).
3. Smart-parking management algorithms in smart city. Scientific Reports (2022).
4. A sustainable strategy for Open Streets in (post)pandemic cities. Communications Physics (2021).
5. Context-aware multi-head self-attentional neural network model for next location prediction. Transportation Research Part C Emerging Technologies (2023).
6. Sidewalk networks: Review and outlook. Computers Environment and Urban Systems (2023).
7. Promoting sustainable urban mobility via automated sidewalk defect detection. Sustainable Development (2024).