+
.webp)
Built on Google Cloud, powered by CARTO
Transcription
Transcript
This customer story has been adapted from a presentation given by Javier Morales Puerta, CTO of the Madrid New Master Plan Office at Ayuntamiento de Madrid (Madrid City Council), at the SDSC26 conference.
Introduction: Madrid at a Glance
Madrid covers 600 square kilometers, 10% of which is protected land, spread across 21 districts and 131 neighborhoods. Nearly 3.5 million people live in the city, rising to almost 7 million across the metropolitan area, alongside 11 million visitors a year. The city has a GDP of €200 billion, and the city council's own budget is close to €6.7 billion.
The Challenge: A Planning Model Stuck in 1997
In Spain, urban planning is governed by general plans — restrictive, all-encompassing legal documents. Madrid's current general plan was approved in 1997. It's rigid, predictive rather than adaptive, and takes 10 to 12 years to approve through a complex process with mandatory reports and limited citizen participation. Litigation levels are high, and by the time a plan is finally approved, circumstances have often moved on. Morales compared the process to the board game of the goose: you can spend over a decade working through it, only for the whole plan to fall apart at the very end.
The rigidity extends down to the smallest details — the 1997 law even regulates the color of building windows. Meanwhile, cities are changing faster than any static, decades-old law can account for.
A New Paradigm: 4D, Data-Driven Urbanism
Madrid's response is a new planning paradigm built around the idea of 4D urbanism — extending traditional 3D urban planning to explicitly account for time. The approach needs to be dynamic, flexible, performance-based, and cross-sector, rather than fixed and siloed.
Crucially, the city is treating this as a shift with three equally important dimensions: urban, legal, and digital. Rather than bolting on a digital tool at the end of the process, the digital component — and citizen dialogue — are built in from the very start. The city has defined 10 challenges every city has to face, including housing, sustainable mobility, and climate neutrality, with citizens part of the conversation from day one.
Introducing the Strategic City Simulator
To put this paradigm into practice, Madrid built the Strategic City Simulator, with an initial version developed in just six months. Rather than forecasting long-term trends, the simulator is designed to recommend actionable decisions for better outcomes today.
It combines demographic, land use, mobility, environmental, and socio-economic data into a single, holistic view of the city — something no single department inside the city council had previously done on its own. Digital simulation now shapes how planning problems are framed and how solutions are designed from the very start of a project, and the simulator itself updates continuously as new data arrives, so the city can monitor, refine, and improve policy over time.
A Data Architecture Built for Trust
The simulator runs on infrastructure built inside the city council's own technical strategy, rather than as an isolated side project — keeping it automatically up to date rather than letting it become another legacy system. The stack is built on Google Cloud, Google BigQuery, and CARTO for visualization.
Governance was treated as a first-class requirement: strong frameworks ensure data quality, security, trust, and traceability. As Morales put it, this is not a black box the city is ruled by — the indicators used to make decisions are published openly, so anyone can see how the city is being run.
From Madrid Inteligente to City Intelligence
The roots of this work go back more than a decade, to the 2014 launch of Madrid Inteligente (Intelligent Madrid), which moved the city from isolated systems to an integrated operational platform now connecting around 300 municipal services across mobility, environment, and energy. That evolved into an open data platform with more than 550 datasets, plus a geoportal distributing the city's geospatial data.
Today, the geoportal alone serves around 180,000 users a month, with 950 datasets and roughly 3.75 million files totaling 80 terabytes of information. This work sits under the city's broader transformation strategy, Madrid Digital Capital, of which city intelligence is one of three strategic pillars.
Measuring What Matters: The CSEC Indicator System
To build the simulator, the city first had to define CSEC, its strategic city simulator indicator system, guided by a principle often attributed to Lord Kelvin: what is not defined cannot be measured, what is not measured cannot be improved, and what is not improved always degrades.
The resulting system draws on 37 traceable public data sources, combined into 134 datasets and 192 indicators — out of 585 potential indicators identified along the way. Each indicator is mapped to one of the city's 10 challenges: housing alone has 46 indicators covering people, rents, housing stock, and age, all visible directly inside the simulator.
Underpinning this is a layered data architecture built for trust and traceability, moving data through three phases: bronze (stored exactly as received), silver (cleaned and normalized into a consistent historical record), and gold (ready to power dashboards, simulations, and decision-support tools).
Multiscale Analysis and Artificial Intelligence
The simulator supports multiscale analysis — from the metropolitan area and individual neighborhoods down to parcels, plots, and blocks — using tools including H3 hexagons.
AI is used throughout, within a transparent, accountable framework, to automate tasks, surface patterns, and support — not replace — the city's own urban planning team. One capability lets planners have a conversation with the maps they build: an AI agent connected to the full data lake can answer questions about a map even when that data isn't visually represented on it, such as asking about rents or age distribution on a mobility map. AI is also used to connect maps back to planning documentation, including the 1997 general plan, surfacing which rules apply to a given part of the city, and to analyze citizen sentiment from social media alongside the city's own indicators.
Putting It Into Practice
Two examples show the simulator in action. The first reprograms parts of the existing general plan around housing: identifying areas to redensify with more residential blocks, generating budget that is reinvested into redesigning the public space around those buildings. The second focuses on green and blue infrastructure — identifying where to plant more gardens and build green streets with additional cycle lanes.
Looking Ahead
Madrid is using the Strategic City Simulator to move from a static, litigation-prone planning model built on a single rigid law to a living, continuously updated system — one where urban, legal, and digital decision-making evolve together, with citizens and open data at the center throughout.
