Road networks are magnificent bird-eye view fingerprints of cities, age-old topics of urban planning, and stable cornerstones of spatial data science. As a primary goal of today’s urban planning is to design livable, future-proof cities via concepts like the 15-minute city, here I collect the top lists of most livable cities and give a visual overview of their road networks – with a ChatGPT twist.
Which are the most livable cities on Earth?
Year after year, numerous rankings are published based on which cities are the most livable, with Western Europe, Canada, and Australia usually topping these lists. These lists try to factor in as many aspects as possible, from public transport to healthcare and urban health. As they intend to be very general,it might very well happen that for certain individuals, even myself, the personalized top list would look different. Diving into this debate may be above my paygrade, so instead, I collected six different rankings credible enough to be included in the City quality of life indices article on Wikipedia.
The lists included in this collection are the following: the Deutsche Bank Liveability Survey, Global Finance’s World’s Best Cities to Live, Mercer’s Quality of Living Ranking, Monocle’s Quality of Life Survey, Numbeo’s Quality of Life Ranking, and The Economist Intelligence Unit’s. All these lists capture somewhat different angles of livability, yet they also show quite some similarities. Aiming at the most generic list of cities possible, I merged them all and picked those cities that were included in at least two of them – which was twelve cities total.
Turns out Zurich is the most popular one here – it appeared on five of the six lists, followed by a tie between Vienna and Copenhagen as each made it to four rankings. Next comes Vancouver, Melbourne, and Sydney which all made it to three of the top lists. Finally, Geneva, Frankfurt, Amsterdam, Helsinki, Tokyo, and Wellington were all double-listed. In the later part of this article, I will refer to these twelve cities as the most livable cities (according to the merged rankings of different but professional opinions).
A color palette from the AI
During the past several weeks, OpenAI’s ChatGPT has been making huge waves in countless domains from AI research to copywriting. As my background is as far from design as Physics, I am not a natural talent with color pallets and such, so instead decided to shoot my shot and ask a few questions from the chatbot like this:
me:
Give me please a color palette of 12 pairs of colors which include the pairs of CherryRed and OffWhite, Coral and Yellow, and BubblegumPink and CherryRed when the colors are ordered based on the primary colors.
chatbot:
Sure, here is a color palette of 12 pairs of colors that include the pairs of Cherry Red and Off White, Coral and Yellow, and Bubblegum Pink and Cherry Red, ordered based on the primary colors:
Cherry Red (#ff0000) and Off White (#f5f5f5)
Coral (#ff7f50) and Yellow (#ffff00)
bubblegum Pink (#ffc0cb) and Cherry Red (#ff0000)
Red-Orange (#ff4500) and Lime Green (#32cd32)
Orange (#ffa500) and Green (#00ff00)
Lime Green (#32cd32) and Turquoise (#40e0d0)
Green (#00ff00) and Blue (#0000ff)
Turquoise (#40e0d0) and Purple (#6600ff)
Blue (#0000ff) and Off White (#f5f5f5)
Purple (#6600ff) and Coral (#ff7f50)
Off White (#f5f5f5) and bubblegum Pink (#ffc0cb)
Coral (#ff7f50) and Cherry Red (#ff0000)
As a reminder, keep in mind that color palettes often depend on the context, the purpose, and the medium where you will use them. Also, this ordering is based on the traditional primary colors but it is not always the case, some palettes could be organize differently.
After receiving a detailed answer in a blink of an eye, ChatGPT even bothered to give me a kind reminder – how thoughtful! After a few different versions of these questions, I ended up settling with two color palettes which I manually merged. The main reason I had to do that was to increase the background-foreground contrast, as the diagrams I have here really need that. While for human eyes, these contrast ratios are probably quite intuitive, it still seemed to be too challenging to properly explain it to the AI – for now.
Finally, Figure 1 shows the pairs of colors generated and matched by the AI, and organized by me.
Urban network visualization
Visualizing and studying urban networks, such as the road network of cities, is a crucial point in numerous spatial data science applications. It comes up in any transportation-related issue from traffic jam forecasting to public transport reconstruction or micro-mobility station localizations and has a major role in accessibility and achieving 15-minute walkable cities as well. Also, they look cool and it’s fun to work with them!
So first, I went for my go-to tool, OSMnx, a Python package that combines the most widely used network science package (NetworkX) and the 101 of spatial data in Python (GeoPandas) by allowing us to download, manipulate, and analyze geospatial data from OpenStreetMap fairly easy.
One of the intro-level functionalities is to download the road network of a specified geometry, such as the administrative boundary of a city. To test out how things work, I used my hometown, an indeed small town of about 10k people in the countryside of Hungary – Siklós. Sikós both has a nice castle and an interesting road network, so the only thing I had left was to combine this network data with the color palette I co-created. Hence, I visualized the inner part of the city, a circle around its centroid with a radius of 1.3km. Additionally, I only used Python, specifically, Matplotlib to create these maps.
Top cities visualized
Finally, I created a gallery of the top twelve most livable cities based on the joint ranking, each visualized using the AI-generated color pairings on the following figures.
With a background in physics and biophysics, I earned my PhD in network and data science in 2020. I studied and researched at the Eötvös Loránd University and the Central European University in Budapest, at the Barabási Lab in Boston, and the Bell Labs in Cambridge. I am currently the chief data scientist of Datapolis, a research affiliate at the Central European University, a senior data scientist at Maven7, and a data science expert of the European Commission.
