You are not a number. You are a human being. But many of your responses to places are predictable, which is good news for anyone trying to create public places for humans, not theories, writes Nicholas Boys Smith.
Thanks to support from Cadogan, Create Streets (the social enterprise I run) has recently been researching the types of spaces people like and the ones they avoid. We reviewed the empirical – not the opinionated – literature.
With MORI, we polled more than 2,000 people with carefully controlled images. We applied a visual preference algorithm – created at the Turing Institute and trained by 1.5m responses to 200,000 images – to 19,000 streets and squares in six British cities. And we conducted a case study of new development near Cambridge station. We wondered whether we could find patterns in the types of places people like by comparing algorithm scores and observed behaviour with the “big data” on our cities. We could.
Start your free trial today
Your trusted daily source of commercial real estate news and analysis. Register now for unlimited digital access throughout April.
Including:
Breaking news, interviews and market updates
Expert legal commentary, market trends and case law
You are not a number. You are a human being. But many of your responses to places are predictable, which is good news for anyone trying to create public places for humans, not theories, writes Nicholas Boys Smith.
Thanks to support from Cadogan, Create Streets (the social enterprise I run) has recently been researching the types of spaces people like and the ones they avoid. We reviewed the empirical – not the opinionated – literature.
With MORI, we polled more than 2,000 people with carefully controlled images. We applied a visual preference algorithm – created at the Turing Institute and trained by 1.5m responses to 200,000 images – to 19,000 streets and squares in six British cities. And we conducted a case study of new development near Cambridge station. We wondered whether we could find patterns in the types of places people like by comparing algorithm scores and observed behaviour with the “big data” on our cities. We could.
The most popular places normally have a high built-up area density, a rich mix of land uses, lots of older and listed buildings, a good sense of enclosure and an above-average diversity of shop types.
“People prefer richly featured buildings with symmetry, coherence, composition and textured materials over blind-glass facades. Urban designers often tell us that buildings don’t matter, “it’s the spaces in between”. They are wrong.”
Some 60% of people prefer to spend time in a medium-sized square, rather than in a larger square. People would rather walk in streets with a height to width ratio of about one-to-one. And, all things being equal, three to five times as many people prefer richly featured buildings with symmetry, coherence, composition and textured materials over blind-glass facades.
Urban designers often tell us that buildings don’t matter, “it’s the spaces in between”. They are wrong.
Using this information, we have developed a list of steps that normally “should” be taken when designing spaces.
1) Gentle density is your friend but “fine grain it”. Three to seven storeys are best, with squares 80m to 100m wide.
2) When it comes to greenery, little and often is normally best. The best front gardens are less than 3.5m deep.
3) Benches and statues should be structured, not randomised. Place benches at the edges of public spaces and away from traffic, ideally with a view.
4) Beauty really, really matters. The most popular places, with a predictable majority, have a strong sense of place and “could not be anywhere”. Carry out some visual preference surveys. Do people say they like it?
5) Mix it up. Have a variety of street types. Encourage self- and custom-build, within a framework that might set height, bay width and some basic rules on materials and facade pattern.
6) Ensure that the edges of buildings in a square appeal both to those walking beside them and those on the other side.
7) People like to feel enclosed, but only up to a point. Build street height-to-width ratios of between 0.75 and 1.5.
8) Walkability works, but does not quite mean maximising space to walk. Design continuous walkable environments that are more than 400m long.
Humans are not entirely pre-programmed but we substantively are. What we like and don’t like is, at heart, a consequence of our humanity and of our contradictory natures, of our need for privacy and for company, for shelter and for stimulation.
That’s why, from Shanghai to Stratford-up-Avon, the best places have many of the same characteristics, why the European tourist enjoys Marrakesh and the Chinese tourist revels in Paris.
That is why, through all the mess of human nature, and through the millions of data points we investigated, we can predict what makes for a popular street or square and what makes for a bad one. Normally.
Nicholas Boys Smith is the director of Create Streets and a member of the Government’s Building Better, Building Beautiful Commission