Computers in Urban Studies

Jon Reades - j.reades@ucl.ac.uk

1st October 2025

All of this may seem very new and very scary if you have a ‘traditional’ background in Planning, Geography, or Urban Studies but…

… we actually have a long history of cutting-edge computer use in these disciplines! Every time there has been a step-change in computing, geography and planning have been there to exploit it!

3 Waves?

1. “A computer in every institution”
2. “A computer in every office”
3. “A computer in every thing”

Wave 1

• Roughly the 1950s–70s
• Computers as ‘super-human’ calculators for data
• Data + models as theory-testing tool

Retrospectively: the 1st quantitative revolution. But can see this as incorrect and focus on the theoretical aspect.

Wave 2

• Roughly the 1980s–2000s
• Computers as tools for thinking about spatial relationships
• Explicit modelling of local spatial effects

Retrospectively: the GIS revolution. But I personally see this as incorrect because GIS is Wave 1.

Shift from computers as processors of data to integrated, pervasive systems that spew out data on everything.

Wave 3

• Roughly the mid-2000s–?
• Computers as tools for generating data (pace ABM researchers)
• Geodata being continuously produced as byproduct of other activities
• Shift from researching attributes to behaviours (pace Hägerstrand)

Retrospectively: the big data revolution or 2nd quantitative revolution.

All Waves Still Going!

Wave 1: Computers help me do it faster

GIS is ‘just’ the industrial revolution impacting cartography.

Wave 2: Computers help me to think

Geocomputation & local stats are qualitatively & meaningfully different.

Wave 3: Computers help me to learn

Not ‘just’ about the ‘bigness’ of data, though that is important.

Wave 2 is about implementing ideas such as recursion and iteration – these could, in theory, have been tackled in Wave 1, but in practice that’s not what people were doing.

Wave 3 is about more explicitly allowing computers to learn about data so that we can extract insight from these models – these could also, in theory, have been tackled in Wave 2 but in practice that’s not what people were doing.

I’m not totally happy about my description of Wave 3 and will try to dig into this in a little more detail but suggestions welcome!

Anticipated by Hägerstrand (1967)

I think that the computer can do three different and useful things for us. The first and simplest operation is… descriptive mapping the second… is the analytical one The third kind of service is… to run process models by which we might try to reproduce observed or create hypothetical chains of events of a geographical nature.

Hägerstrand (1967)

But Persistent Critiques

More important, there is a clear disparity between the sophisticated theoretical and methodological framework which we are using and our ability to say anything really meaningful about events as they unfold around us. There are too many anomalies between what we purport to explain and manipulate and what actually happens. There is an ecological problem, an urban problem, an international trade problem, and yet we seem incapable of saying anything of any depth or profundity about any of them. When we do say something it appears trite and rather ludicrous.

Harvey (2008 [1972], 17)

Data science?

Source: Davenport and Patil (2012)

• Industry-led
• Spatially ignorant (often)
• Disciplinarily greedy (often)

Source: Schneider (2013)

Personally, I think that you get to danger zones everywhere here, not just at the intersection between hacking and substantive expertise.

What is Different?

According to Donoho (2017) ‘data science’ differs from plain old ‘statistics’ through an interest in:

1. Data gathering, preparation, and exploration;
2. Data representation and transformation;
3. Computing with data;
4. Data visualisation and presentation;
5. Data modelling; and
6. A reflexive ‘science of data science’.

In Practice…

I think there are several distinguishing features that I encounter in day-to-day (geography) work:

• Data-driven methods development & deployment
• Explicit tuning/meta-parameterisation
• Explicit feature optimisation/engineering
• Explicit training/testing from ‘one shot’ data
• ‘Black boxes’ feature prominently & ‘online learning’ emerging quickly

Data science as process and pipeline, not just input to research.

Do We Need It?

Source: xkcd

Additional Resources

• The Data of Urban Spaces — a nice collection of pieces on Towards Data Science

Thank You

References

Davenport, T. H., and D. J. Patil. 2012. “Data Scientist: The Sexiest Job of the 21st Century.” Harvard Business Review. https://hbr.org/2012/10/data-scientist-the-sexiest-job-of-the-21st-century.

Donoho, D. 2017. “50 Years of Data Science.” Journal of Computational and Graphical Statistics 26 (4):745–66. https://doi.org/10.1007/978-3-642-23430-9_71.

Hägerstrand, Torsten. 1967. “The Computer and the Geographer.” Transactions of the Institute of British Geographers. JSTOR, 1–19.

Harvey, D. 2008 [1972]. “Revolutionary and Counter-Revolutionary Theory in Geography and the Problem of Ghetto Formation.” In Geographic Thought, edited by G. Henderson and M. Waterstone, 1st Edition. Routledge. https://doi.org/10.4324/9780203893074.

Schneider, A. 2013. “How to Become a Data Scientist Before You Graduate.” Berkeley Science Review. http://berkeleysciencereview.com/how-to-become-a-data-scientist-before-you-graduate/.