EN TR
Urban Metabolism Research
Context & Simulation07 - CITYWorking Note2024

Urban Metabolism

Cities breathe, circulate, and decay. We spent twenty years of public data learning to read the vital signs.

Scale 07 City
Cities 5 compared
Data Span 20 years
Tools GIS Python
Type Working Note
Updated 2026-08
01

A city is not a machine. It is an organism.

The metaphor has a scientific lineage. In 1965 the sanitary engineer Abel Wolman formalized 'the metabolism of cities': treat a city as a living system with inputs, flows, and wastes, and its health becomes something you can measure. Christopher Kennedy and colleagues turned that idea into an accounting method for world cities in 2007. And Michael Batty and Paul Longley showed in Fractal Cities that healthy urban growth is self-similar, repeating its patterns from the block to the region. That last idea is close to home: it is where our name comes from.

We asked a narrow version of their question: which measurable flows separate cities that work from cities that merely grow? We read five cities through the organism lens: Tokyo (density carried by rail), Copenhagen (human scale as policy), Istanbul (organic continuity), London (the industrial original), and Ankara, the city we live in. This is a desk study: public records, GIS layers, operator statistics, and twenty years of documentation, plus the one source no dataset provides: living here.

The starkest signal is circulation. Greater Tokyo's rail web carries about 40 million journeys a day; per resident, that is roughly 13 times more rail use than Ankara's network manages. Since 2007 Ankara has added about 1.4 million residents while its rail lines grew from 23 to 67 kilometers, and the strain is documented: in August 2007 the city's reservoirs fell below five percent of capacity, and in 2025, with dam levels near twenty percent, supply interruptions returned.

One honesty note before the charts. Every number on this page is either compiled from a named public source or is our own scored index, and we flag which is which. Cross-city figures mix operators and definitions; read them as orders of magnitude, not decimals.

Comparison panels of urban metabolisms

Metabolisms Compared: Comparison panels from the five-city desk study: rail intensity, street activity, and neighborhood identity.

02

Theoretical Framework

01

Circulation

How people and goods move: rail, buses, bike lanes, streets. Our proxy: daily rail trips per 1,000 residents. Greater Tokyo scores about 1,080; Ankara about 85. When circulation fails, everything clogs.

02

Nervous System

The informal encounters that make urban life rich. Neighbors chatting, strangers nodding, kids playing. Our proxy: the share of streets with regular pedestrian activity, counted the way Jan Gehl taught.

03

Organs

Distinct neighborhoods with their own character. Shibuya feels different from Shinjuku; Kadikoy feels different from Besiktas. Our proxy: a scored district-identity index. Healthy cities have organs. Sick cities are uniform.

04

Skin

The boundary between private and public. Active ground floors, permeable facades, eyes on the street. Our proxy: the share of street frontage with active ground-floor use.

05

Immune System

Planning and regulation. Does infrastructure precede settlement, or chase it? Plan-led growth is a working immune response. Demand-led sprawl is the disease arriving before the antibodies.

03

Research Process

01

Define the Organism

We established five biological systems to read: circulation, nerves, organs, skin, immune response, and paired each with one measurable proxy

02

Gather the Data

TUIK statistics, EGO and Tokyo operator ridership, municipal bicycle accounts, GIS layers, and satellite classifications across five cities

03

Map the Systems

Translated urban data into metabolic indicators. Rail intensity became arterial health. Street activity became neural activity.

04

Compare and Diagnose

Cross-referenced cities to identify what separates healthy urban organisms from struggling ones

04

Research Phases

01

City Selection

We chose five cities representing different urban philosophies: Tokyo (density), Copenhagen (livability by policy), Istanbul (organic growth), London (industrial legacy), Ankara (sprawl).

02

Data Collection

Twenty years of public data: TUIK population statistics, EGO and Tokyo operator ridership, GIS layers, satellite classifications of built-up area. For Ankara we added the one source no dataset provides: living here.

03

Metabolic Mapping

Translating raw data into biological readings. Rail intensity as arterial health. Pedestrian zones as active tissue. Dead zones as necrosis. Each system paired with one measurable proxy.

04

Comparative Analysis

What separates thriving cities from struggling ones? We cross-referenced the five cities and identified four indicators that track urban health.

Mapping diagram: organism systems, their urban counterparts, and measurable proxies

The Reading Frame: Five organism systems, their urban counterparts, and the proxy we can actually measure for each. The metaphor only earns its keep when every row has a number.

05

Key Metrics

1,080
rail trips / 1,000
Greater Tokyo, daily (operator statistics)
85
rail trips / 1,000
Ankara, daily (EGO, 2025)
62%
bike commutes
Copenhagen residents, work and study trips (2021)
<5%
reservoir volume
Ankara, August 2007
06

Key Thinkers

01

Abel Wolman

American Sanitary Engineer, 1892-1989

Coined 'the metabolism of cities' in a 1965 Scientific American article: a city as a system of inputs, flows, and wastes. Every urban-metabolism study since, including this note, works inside his frame.

02

Jan Gehl

Danish Architect and Urban Designer

Gehl spent fifty years counting pedestrians. His Copenhagen street studies proved that walkable cities are not just pleasant, they are economically and socially essential. Our street-life scoring builds directly on his counting methods.

03

Jane Jacobs

American-Canadian Activist and Writer

In 1961, Jacobs showed that healthy neighborhoods need mixed use, short blocks, and old buildings. Tokyo has all three. Ankara's new districts have none.

04

Michael Batty

British Urban Scientist, CASA UCL

Fractal Cities (1994, with Paul Longley) showed that organic urban growth is self-similar from block to region. The New Science of Cities rebuilt planning on networks and flows. The lens this note borrows, and the idea our studio is named after.

07

Case Studies

Ankara: The Full Autopsy

Turkey

From 2007 to 2024 Ankara added roughly 1.4 million residents (TUIK) while its rail network grew from 23 to 67 kilometers. Satellite classifications from 1988 to 2020 show the built footprint expanding outward faster than any network that serves it. The strain is documented: in August 2007 reservoir volume fell below five percent of capacity and a 130-kilometer emergency pipeline to the Kizilirmak was laid in eleven months. In 2025, with dam levels near twenty percent, supply interruptions returned.

+1.4M Population Growth
23→67 km Rail Network
<5% Reservoirs, 2007

Tokyo: The Functioning Giant

Japan

37 million people in one metropolitan area, yet it works. The secret is density carried by rail: 304 kilometers of subway inside a regional rail web of roughly 2,600 kilometers moving 40 million journeys a day. Trains every two to three minutes at rush hour, and neighborhoods that kept their identities through modernization. Shibuya, Shinjuku, Ginza: each feels like a different city.

37M Population
40M Daily Rail Trips
92 (our index) District Identity

Copenhagen: The Human Scale

Denmark

Europe's most livable city did not happen by accident. In 1962 Copenhagen closed Stroget, its main street, to cars. Merchants predicted disaster; pedestrian traffic rose about 35 percent within the first year, and the city spent the next four decades converting street space to people space. Today 62 percent of residents' trips to work or study are made by bicycle. The result: cleaner air, lower transport costs, and streets where people actually talk to each other.

1962 Car-Free Since
62% Bike Commutes
+35% Ped. Traffic, Year One

Comparative Analysis

Illustrative study model: hyper-dense metropolis carried by layered rail

Tokyo

Complexity That Works

23 named wards, each with distinct character. A 304-kilometer subway core inside a 2,600-kilometer rail region. Trains every 2-3 minutes. Villages merged into the city but kept their names and identities.

37M peopleDenseRail-rich
Illustrative study model: low-rise harbor city with a pedestrian spine

Copenhagen

The Human City

62 percent of resident commutes by bike. Stroget, one of Europe's longest pedestrian streets, car-free since 1962. Green space within reach of most homes. Small, intentional, livable.

2M peopleWalkableGreen
Illustrative study model: organic peninsula city beside a strait

Istanbul

Organic Survival

More than 2,500 years of continuous habitation. Byzantine street patterns still in use. The Bosphorus naturally limits sprawl. Chaotic but alive.

15M peopleHistoricGrowing
Illustrative study model: fragmented sprawl drifting across a plateau

Ankara

The Warning

67 kilometers of metro for 5.5 million people, and roughly 13 times fewer rail trips per resident than Tokyo. Infrastructure built after the fact, if ever. This is what not to do.

5.5M peopleSprawlingCar-dependent
05

Optimization Results

100% 75% 50% 25% 0%
1080%
520%
330%
150%
85%
Tokyo
London
Copenhagen
Istanbul
Ankara

How often does the average resident actually board a train? Network length flatters; ridership tells the truth.

Compiled from operator statistics (2024-2025), metro-area denominators. Order-of-magnitude comparison.

08

Key Findings

01

Circulation predicts almost everything. Greater Tokyo residents make roughly 13 times more rail trips per person than Ankara residents. Across our five cities, the rail-intensity ranking tracks the livability ranking almost exactly.

13× rail gap
02

Street life is measurable. Copenhagen scores 60 percent active streets in our count, Ankara 15 percent. Dead streets do not just feel bad; Gehl's half-century of counts links them to isolation and distrust.

4× difference
03

Neighborhood identity matters. In Tokyo, each ward has a name, a personality, a reason to exist. In Ankara's new districts, everything looks and feels the same. Anonymity breeds alienation.

92 vs 28 (our index)
04

Infrastructure-first works. Copenhagen's Finger Plan (1947) and Tokyo's rail-led growth put the network before the buildings. Ankara built the buildings first; the network is still catching up two decades later.

Plan-led vs demand-led
09

Honest Limitations

Data Dependency

This is a desk study. We compiled and scored public data; we did not run field measurements in five cities.

Data Dependency

Cross-city numbers mix operators, network definitions, and denominators. We treat them as orders of magnitude and flag every figure that is an estimate.

Data Dependency

Five cities is not a large sample. We see patterns, but we cannot claim statistical certainty.

Data Dependency

Our indices for street life and neighborhood identity involve judgment calls. Others might score differently.

Temporal Limitation

Twenty years of Ankara data is substantial, but urban change happens over centuries. We are seeing a snapshot.

Behavioral Assumption

Istanbul works despite chaos. Some cities break the rules and thrive. We do not fully understand why.

10

Where This Led

This study shaped how we think about district context. The metabolic framework informs the context-analysis direction we're exploring for Archly, and seeded a set of Ideal City planning assets now in preparation: 3D models drawing on Copenhagen's density and Tokyo's transit logic.

Archly Product
Metabolic Analysis Method
11

Conclusion

Cities are organisms. They have circulation, nervous systems, organs, and immune responses, and each of those systems can be paired with a number you can actually check. When the systems work together, urban life thrives. When they do not, cities become collections of buildings where no one wants to live. Ankara taught us what strain looks like. Tokyo and Copenhagen showed us what circulation and street life can do.

Limitations

  • Five-city sample
  • Desk study, not fieldwork

Future Directions

  • Extend the reading beyond five cities
  • Publish the compiled dataset
  • Field-verify the street-life index in Ankara