Data-Driven City Lecture Notes

Urban Challenges

• Urbanization: Cities centralize $70\%$ of the world's GDP and over $50\%$ of the population (2017 data).
• Resource Usage: Cities use $75\%$ of natural resources.
• Emissions: Cities produce about $70\%$ of greenhouse gas emissions.
• Population Growth: Urban population expected to increase from $3.6$ billion to $6.3$ billion by 2050 (United Nations, 2018), making two-thirds of the global population urban.
• New Problems: Traffic congestion, waste management, pollution, and parking allocation.
• Limited Resources: Adapting the city to current and future needs is a priority.
• Interdisciplinary Collaboration: Professionals from various disciplines are collaborating to create smart cities.
• Intelligence Measurement: Questions arise on how to measure a city's intelligence.
• Urgency: Initiatives worldwide are addressing the need to make cities more suitable.

Environment and Quality of Life

• Environmental Issues: The need for greener cities.
• Resource Consumption: Cities consume $75\%$ of overall resources but occupy only $2\%$ of the land.
• Greenhouse Emissions: Cities account for $70\%$ of greenhouse emissions (World Bank, 2017).
• Climate Change: Cities are affected by climate change (e.g., Venice, Tokyo, Amsterdam, Shanghai, Jakarta).
• Air Pollution: $8.8$ million premature deaths worldwide due to air pollution.
• Quality of Life: Related to economic conditions.
• Poverty: Leads to exclusion from society and social systems (World Bank, 2018).
• Housing Problems
• Criminality and Security: Both real and perceived.
• Human Mobility, Freight Transportation, and Logistics

Sustainability and Resilience

• Sustainability:
  • Economic: Foster entrepreneurship and local production.
  • Environmental: Waste management and recycling, local food production, and land use.
  • Civic/Social: Foster participation to create people-oriented smart cities.
• Resilience:
  • Definition: The ability to prevent adverse events, cope with them, and adapt successfully.
  • Examples: Terrorist attacks (Paris 2015), Katrina storm, COVID-19, Cyberattacks.
  • Data Management: Correct data management and resilient communication systems are key for resilient value chains and urban services.
  • Integration and Coordination: Planning for integration among city components is crucial.

Smart, Intelligent, Digital City

• Concept: Using information and communication technologies to improve city functioning.
• Synonyms: Intelligent city, Digital City.
• Perspective: Often biased towards an ICT-centric view.
• Definition Issues: No universally accepted definition.
• MIT Definition: Systems of systems with digital nervous systems and intelligent responsiveness.
• World Bank Definition: Adopting technical and information platforms to better manage resources, improve management, monitor developments, develop new business models, and help citizens make informed decisions.
• Effective Integration: Integration of physical, digital, and human systems for a sustainable, prosperous, and inclusive future.
• Smart City Defined: When social capital and modern information and communication infrastructure fuel sustainable economic development and high quality of life.
• Enhancement of urban services, reduction of costs and resource consumption, and effective citizen engagement.
• Innovation and fulfilling needs are central.

Priorities of a City – A Roadmap

• Prioritization: After finance, prioritization is the second most significant barrier to smart city strategy implementation (PWC Survey).
  1. Basic Needs: Security, health, and basic infrastructure (clean water and sanitation).
  2. Infrastructure: Roads and transport, educational access as the city evolves from basic industrial production to an informational society.
  3. Environmental and Social Needs: Environmental needs, social integration, culture and leisure, with ICT enabling a knowledge-based society.
  4. Smart City: Becoming a world leader to maximize performance across all capitals.
  5. Self-Actualization: Exploring new paradigms and setting new standards for quality of life, sharing experiences to help other cities advance.

Pillars and Enablers

• Smart Living: Quality of life aspects like culture, health, safety, housing, and tourism.
• Smart Societies: Agile civil society, education, social inclusion, quality of social interactions, active participation, and openness.
• Smart Environment: Attractive natural conditions, resource management, and pollution levels.
• Smart Economy: Economic competitiveness factors such as innovation, entrepreneurship, trademarks, productivity, and integration in the international market.
• Smart Governance: Political participation, services for citizens, and administration functioning.

Smart City as an Opportunity for Innovation

• Economic Drivers: Cities drive individual state economies.
• GDP Contribution: Paris covers $30\%$ of French GDP with only $16\%$ of the population.
• Education Levels: Generally offer high levels of education.
• Market Access: Large customer market in a compact area.
• Start-up Hotbeds: Cities foster start-ups.
• Urban Entrepreneurship: New type of entrepreneur focusing on city-specific services.
• Service Examples: Uber, Moovit.
• Open Data: Used by entrepreneurs to generate value and return services.

Smart City as a Sustainable City

• Objectives: Increasing human well-being and fostering economic growth.
• Constraints: Scarce natural resources and cultural barriers.
• Urban Planning: Rational use of urban land and conscious design of built environment.
  • Compact City: Work, live, and leisure in the same neighborhoods to shorten commutes.
  • Eco-City: Minimize resource input and waste output, design green infrastructures, and implement pervasive waste recycling.
• Sustainable City (Bibri and Krogstie, 2017):
  • Maximizes energy and material use efficiency.
  • Zero-waste system.
  • Supports renewable energy-based production and consumption.
  • Promotes carbon-neutrality.
  • Provides sustainable mobility services.
  • Preserves the whole ecosystem.
  • Fosters livability and community-based human conditions.
• Inclusive digital access to prevent technological divides.
• Privacy protection and ethical data management frameworks.
• Enhanced social cohesion using technology to strengthen community bonds.
• Supporting digital platforms that connect neighbors for resource sharing.
• Creating mixed-use smart spaces that encourage diverse social interaction.
• Equitable distribution of services across all neighborhoods.
• Cultural preservation that maintains local identity alongside modernization.

ICT Contribution to Environmental and Social Sustainability

• Lights and Shadows:
  • Fast Development Cycles: ICT consumer market leads to high volumes of e-waste and emissions due to transportation from Asia.
  • Digital Divide: Consequences on employment and access to city’s digital opportunities.
  • Privacy and Security: Important discussion points.
  • Data Access: Considerations on who decides and who gets access to the data.
  • Trade-off: Balance between security and privacy.
• Resource Optimization: IoT, Cloud, Big Data, and AI can optimize resource use (doing more with less).
• Shared mobility, smart grids, water management, optimized waste collection.
• Circular economy via mobile apps.
• Work-from-home

Pillars and Enablers

• Smart Mobility: Effective and efficient mobility system using innovations in ICT and travel means.
• Smartness: Includes monitoring, control, and optimization for efficiency, bottom-line benefits, and environmental improvements.
• Enablers: Digitalization, automation, connectivity, and analytics for monitoring, control, and optimization.
• Automation - Enables technological infrastructure for smart systems to operate with minimal human intervention.
• Connectivity - Provides communication networks and infrastructure that connect different systems and allow them to interact seamlessly.
• Analytics - Offers data processing capabilities that transform raw information into actionable insights.
• Digitalization - Converts analog processes into digital ones, making governance more efficient and accessible.

Main Elements of the Smart City

• Technology: Sensors, networks, and other technologies to collect data and monitor city operations.
• Infrastructure: Transportation and energy systems to support city operations.
• Governance: Policies and regulations to manage city operations.
• People: Active citizen participation in governance.

Taxonomy of a Smart City

• Technology-oriented approach:
  • Technology's role: Optimize natural resources, energy distribution, and transportation.
  • Data flow: Unidirectionally from city to decision-makers.
• People-oriented approach:
  • Citizen access: Encourages access to data and citizen participation in initiatives.
  • Data flow: Data produced by the city goes to the citizens.

Smart City as an Augmented City

• Enablers: Big data, analytics, and machine learning promise to disrupt city understanding
• Data-driven decision making: Policies should be less subjective and more data/algorithm-based.
• Data Collection: Public and private entities collect data from CCTV, Cadaster, cell phones, meteorological stations, smart traffic lights, and GPS.
• Open Data: Fosters transparency, accountability, and entrepreneurship.

Data-Driven City

• Technology is a mean not the goal.
• Vision and purpose: Tech must have a clear vision benefiting city management and citizens.
• Data Focus: Tech must focus on obtaining data for better city behavior knowledge.
• Data Interpretation: Data needs correct interpretation.
• Data-driven approach:
  • Data collection: (Geolocalized) Data collection (IoT, Edge Computing)
  • Data transportation: (5G, low power wireless networks)
  • Data storage: (Data Lake, SQL/Cloud based solutions…)
  • Data preparation: (Big Data platforms)
  • Data analytics: (Machine learning/deeplearning/optimization/simulation)
  • Data visualization: (Dashboards and control panels)

Urban Informatics

• Definition: Application of computer science to urban environment problems and the technological backbone of a smart city project.
• Domains (Foth et al., 2011):
  • Places: Understanding urban spaces.
  • People: Core element of urban analyses.
  • Technologies: Supporting analyses, from data collection to analytics.
• Definition: Study, design, and practice of urban experiences created by real-time technology and augmentation that mediates physical and digital layers.
• Information processing: From various sources, including resource control via sensor networks and individual's interaction with technologies.
• Applications:
  • Mobility
  • Energy
  • Facility management
  • Healthcare
  • Security
  • Governance
  • Economy

Smart Mobility

• Improvement Areas: Large room for improvement in traditional urban mobility.
• Problems: Car-centered city with inefficiencies, traffic, and air pollution; mobility poverty and non-inclusive mobility.
• Solutions:
  • Smart parking
  • Shared mobility (bikes, e-mopeds, scooters)
  • Mobility as a service – Mobility on Demand
  • Cashless/Digital payments systems
  • Real-time traffic monitoring via CCTV and AI
  • Integrated services (public and private)

Smart Mobility and Analytics

• Examples:
  • Real-time video analytics for traffic monitoring/forecast.
  • Vehicle – people tracking (vision).
  • Parking capacity estimation using satellite images (vision).
  • Traffic prediction (spatiotemporal).
  • Autonomous driving (vision).
  • Smart traffic lights (vision + optimization).

Smart Energy

• Smart Grid: Bidirectional, with energy produced by users.
• Local production – Energy communities.
• Energy sharing Optimization algorithms.
• Smart Meters: Collects data about energy demand, supply, and usage patterns.
• Reduce non-renewable energy.
• Customize tariffs to incentivize to use off-peak energy.
• Smart Grid Definition: An intelligent electricity network integrating multiple power sources.
  • Solar Cells
  • Wind Turbines

Smart Facility Management and Smart Buildings

• Aim: Deliver support services effectively to the built environment.
  • Health
  • Safety
  • Energy Consumption
  • Security
• Smart Buildings: Equipped with sensors to understand usage patterns and optimize resource use.
  • Smart lights and heating
  • Air quality management systems
  • Intrusion detection systems
  • Shared facilities as in co-working environments

Urban Digital Twins for Smart Governance

• City Digital Twin: Develop tools for test-bedding concepts, planning, and decision-making, considering correlations between urban concerns.
  • Simulation
  • What-If analysis
  • Mirroring
• Data-informed model of a complex system, such as a city.
• Virtual representation of physical/digital elements and urban flows.
• Combination of isolated models and data flow.
  • 3D city model
  • Electric network
  • Water pipe network
  • Road network

Leveraging Data Science in Smart Cities

• Definition: Data Science is not only about processes and techniques for understanding our world through data analysis but also decision-making.
• Promise: Provide decision makers with the tools for making informed decisions.
  • Visualizations
  • Statistics and probability
  • Simulations
  • Optimization
• Data science appealing today due to:
  • Deluge of data
  • Available in large quantities and at relatively low cost.
  • Computing power
  • The evolution of hardware, distributed software systems, and Cloud and Edge computing services.

Problems Addressable with Data Science

• Classification problems: estimate the probability of a given observation to belong to a set of classes.
  • Identify citizens at risk to fall below the poverty line in order to offer a subsidy.
• Value estimation problems: find a specific number for a given observation.
  • Estimate real estate values.
  • Estimate the number of employees to deploy a given policy in a neighborhood.
• Similarity matching problems: identify similarity patterns in different observations.
  • Recommendation systems of city services/activities based on similarity scores.
• Clustering problems: grouping observations based on their similarity.
  • Area clustering to propose different tariffs for the public transportation system.
• Association problems: identification of co-occurrence of events.
  • Correlate the impact of public events on the vehicular traffic.

Spatial Data Science

• Definition: Intersection of Geographic Information System (GIS) with Data Science to interpret spatial data.
• Main aspects: geography, (spatiotemporal) statistics, and computer science.
• GIS: System for creating, managing, visualizing, and analyzing data and its spatial component.
• Application: GIS and Data Science (statistics, learning and decision-making), are used to solve geographically oriented problems with focus on statistical analysis of patterns.
• Insights: By exploiting availability of urban big data and new technologies, it is possible to capture insights and knowledge useful for decision-making processes.
• Spatial analysis change: Leading to a change in the way spatial analysis is done, with a greater focus on programming rather than traditional software
• Data science focus: Spatial data science can be viewed as a subset of generic "data science" that focuses on the special characteristics of spatial data, i.e., the importance of "where."
  • Spatial interaction, treats location, distance, as core aspects of the data and employs specialized methods and software to store, retrieve, explore, analyze, visualize and learn from such data

What is spatial analytics?

• Much more than points on a map.
  • (Spatial) cleaning, manipulations, transformations, feature engineering and application of analytics on spatial data.
  • (Spatiotemporal) knowledge discovery.
  • Mobility/parking pattern discovery.
  • Temporal and spatial correlation in traffic times.
• Spatial analytics questions (from prof. Luc Anselin)
  • Where do things happen: cluster, patterns, hot spots.
  • Why do they happen where they happen: location decisions, points of interest.
  • How does the context affect what happen: interactions, correlation.
  • Where things should be located: optimization.
• Spatiotemporal analytics if:
  • Geo-spatial data with timestamp (location+timestamp+value).
  • Question: how time affect where, why, how and should?

First and foremost, indicators

• An indicator is a number that can be interpreted and compared to provide information on patterns or trends, which will ultimately allow us to make more accurate decisions
  • Distilling information.
  • Understanding a phenomenon.
  • Making informed decisions.
• What are the relevant pieces of information?
• City data are rarely static, meaning that they must have historical depth (multivariate time series) for retrospective and prospective studies.
• City data are always geospatial data, information is related to different areas with different positions in the space.
• Rarely there is a single indicator able to grasp all the facets of a certain phenomenon.

Tools… of course.

• Visualizations and dashboards, extremely important because most decisions are not automated or suggested but still inspired.
• Statistics, ML and DL. Several tool for Spatiotemporal data science are available.
  • Folium, GeoPandas, MovingPandas, Qgis, PostGIS.
  • LibCity, a library collecting DL algorithms and dataset for traffic related problems.
• Simulation.
  • SUMO (Simulator for Urban Mobility)
• Optimization
  • Classical approaches based on linear and nonlinear optimization models alternate with bio-inspired or statistics-based approaches (Bayesian Optimization) and reinforcement learning.