At least for Europe it is obvious: all roads lead to Rome! You can reach the eternal city on almost 500,000 routes from all across the continent. Which road would you take?
To approach one of the biggest unsolved quests of mobility, the first question we asked ourselves was: where do you start from, when you want to know every road to Rome? We aligned starting points in a 26,503,452 km² grid covering all of Europe. Every cell of this grid contains a starting point to one of our journeys to Rome.
Now that we had our 486,713 starting points we needed to find out how we could reach Rome. For this we created an algorithm that calculated one route for every trip. The more often a specific single street segment was used, the stronger it is displayed on the map. The maps as an outcome of this project are somewhere between information visualization and data art, unveiling mobility on a very large scale.
The saying "all roads lead to Rome" might refer to a bronze monument built in 20 BC called Miliarum Aureum (lat. "Golden Milestone"). This milestone was located in Rome and erected by the Roman Emperor Caesar Augustus. The Miliarum Aureum was used as a reference point for travelling throughout the Roman Empire. It was considered that all roads lead to this monument.Read more
Now that we had our 486,713 starting points we needed to find out how we could reach Rome. For this we created an algorithm that calculates one route leading to Rome, Italy, for every starting point in Europe. The more often a specific single street segment was used, the stronger it is displayed on the map.
The calculation and presentation of these different routes is based on different pieces of open source software. GraphHopper was our central tool for routing our journeys. In total GraphHopper ran 20 hours while calculating all the routes displayed on the maps. GraphHopper works with Open Street Map data, which has also been essential for our project.
The calculation and presentation of these different routes is based on different pieces of open source software. GraphHopper was our central tool for routing our journeys. In total GraphHopper ran 20 hours while caluculating all routes displayed on the maps. GraphHopper works with Open Street Map data, which has also been essetial for our project.
To find all roads to Rome it took GraphHopper more than 5 hours (on a MacPro), calculating 486.713 Routes from every point in Europe to Rome.
Maybe all roads lead to Rome, because there is a Rome on every continent? During further research we found out that there actually is a city called Rome (or Roma) on every continent of the world. The United States of America actually has 9 cities named after the Italian capital.
Accepting this fact, we adjusted our routing to find the closest Rome to every location in the USA. Each color represents the routes leading to the closest Rome from the starting points. Adjusting and coloring the routing to multiple destinations resulted in a very interesting territorial picture. Thus, every location is connected to the nearest Rome according to fastest travel time.
As we know the melting pot is the home of many Americans who centuries ago migrated from Europe. Needless to say, cities with the name of Rome must have something to do with Italy.
A way of interpreting the spatial pattern of Romes in the United States is looking at migration statistics from the settlement of the New World - especially those immigrants originating from Italy. The map "Besiedlungsgang in den USA" gives a great reference as to when, from where and to where, European settlers came to the United States.
You might have seen an earlier version of this map showing 10 Romes in the United States. This turned out to be a mistake in our data processing. We assumed that this was a city located on the border of two states just like Ulm and Neu-Ulm in Germany. Thanks to an attentive user we've double checked and found that there are actually only 9 Romes - sorry. To find all routes to the 9 Romes of the United States our algorithm calculated for about two hours and found 312.719 routes.
Following this idea of the shortest travel times from any given location, the next step of our mobility quest lead to the capital city of each state in the USA. What territory do the routes to each state capital cover? What is the most remote place from the state capital? What kinds of road networks are vizualized in different parts of the country?
When looking only at service areas, it seems that many states keep their territorial shapes. The coastal states are especially recognizable. The southern states seem to lose their original form.
Also remarkable are the different road networks throughout the States. While the east coast seems to rely on roads parallel to the coast, the Rockies and Appalachian mountains reveal their topography with their curvy roads containing some blank spots not reachable by any road. While the mid-western states of the USA show long straight road networks in rectangular alignment.
To find the shortest route to every capital city in the United States took GraphHopper 1.5 hours, resulting in 290,573 routes.
Realigning territories by travel time also works in Europe. "New Europe" shows big changes to the political borders as we know them today - assuming they correspond with travel-time zones (country capitals).
Some countries stay more or less the same, like Turkey, the UK, Denmark, Spain and Portugal. While tiny states like Andorra, Lichtenstein, Vatican and Monaco experience huge spatial growth. Showing the importance of their infrastructural inclusion.
Looking at the states of central Europe it can get very hard to recognize any of the known political borders. Especially Germany, Switzerland and Italy, amongst many others, seem to lose their characteristic borders.
What happens if we change the saying to: all roads lead to Paris, or Berlin? Walking down the political road, we tried to find out if our mappings could reveal the political structure of a country. We chose to compare Germany and France.
This time all routes lead to the respective capital of a single country, resulting in different spatial arrangements of the most important roads used. Assuming everybody in France and Germany decided to travel to their nation's capital city, the location of traffic jams would be obvious!
Keeping the political constructs of both countries in mind, you can also try to find centralized or federalized structures in the routing visualization. Routes in France seem lead to Paris rather directly, with many heavily used roads. The situation in Germany suggests a rather broken down network of roads, before getting to Berlin.
Another question that was raised was, what would happen if the cities had a different location. Trying to simulate this by simply using another city is nearly impossible. Since the infrastructure of a country is lead by political decisions, it involves a long history of hierarchical decisions when building new roads, etc.