Feeds

Urban Morphology

Both co-founders of the UDL have an original background in Urban Design and continued research in academia, where they independently did research to understand the built environment and how to use it in modern data science to model urban behavior, such as household location choice. 

Example given: when your household decides to move to a new city, you will have to select and compare amongst available apartments on the market. But you don’t know the city. So beside the size and age of the apartment - how do you decide which one to take? You go there and see if you like a specific neighborhood or location. And what exactly do you see and like or dislike? Is it the average income, age of households and origin of households in the neighborhood or number of jobs and types of jobs? Or is it what you see and experience when walking around - the inner courtyard, the wide street, the narrow paths to reach a bus stop? Funny enough these things are often related to each other - because households have preferences in location choice which again drives functions such as bars to react to existing situations. So why do we not use the actual structure of cities in our predictive analytics?

This guided our research! We created tools, developed methods, wrote papers, and came to the same conclusion - there is an incredible high potential and need to extract machine readable information on the structure of cities. That is where the mission of UDL began.

These days, we use bespoke tools to look at the shapes of buildings, spatial distribution on blocks, connectivity of street networks and dozens of other characteristics to generate a unique set of insights into the built environment we all inhabit. And we see that how the city is formed can tell a lot about people living in its different parts, consumer behavior on different streets, or demands for services, such as retail. So, we use it. Our predictive models are based on morphology, making them widely applicable even in cases when thematic data are unavailable , too sensitive or too expensive.
You can check the maps attached to this post to make it a bit less abstract. Here we have six morphological attributes in Zürich, extracted from our Swiss-wide database, each different from the previous one, each encoding different information about the city. And each being able to be used in machine learning prediction such as real estate prices. 

Do you have questions about urban morphology? Want to test it in your software or project? Or want to see it applied outside of Switzerland, in your city? Let us know. We are happy to help!
Made on
Tilda