The energy strategy of the Swiss government foresees a major shift away from nuclear power plants to rooftop solar. The plan has been criticized as unrealistic by some. But we, and probably most of our readers, were missing any intuition to judge whether the plan is realistic. We wanted to _see_ how Swiss cities will have to change to become giant solar power plants. What would Swiss cities look like in 2050? How many of our neighbors would have to install rooftop solar next year? Have a look at our model city “Sonnstetten” and see for yourself.
The article has been in the top 5 percent of our most-read articles of the year. It has received the predicate “wonderful” from a Swiss Federal Energy Agency representative. A seasoned journalist said he had not read such a clear exposition of the topic before.
**Modelling a Swiss city:** Because we wanted the illustration to be close to reality, we decided to use part of a real city (Zürich Altstetten) as the basis for our model. A 3D model of the rooftops exists, which we have edited in blender.
**From terawatt-hours to rooftops:** the data we had were “terawatt-hours per year”. The question we had to answer was how many roofs should receive solar cells in a given year. For this, we based ourselves on a study that estimated the percentage of roofs needed in 2050 and projected backward from there.
**Which roofs should be considered?** We knew from studies that about 60 percent of all roof areas are not well-suited for installing photovoltaics. We used geodata indicating the solar potential of each roof to exclude poorly-suited roofs. This and some background knowledge on solar installation let us make informed decisions on which rooftops would be most likely to get solar panels, year by year.
**Panoramic images on mobile:** the image of the model city was in landscape format but most readers on mobile devices would hold it in portrait mode. So we developed an interactive that lets readers move their phones left and right to explore the whole panorama.
**Unfamiliar interactions:** this interaction is not common. Therefore we conducted user tests to find out where they would stumble and help us refine the design. Also, multiple feedback rounds were needed to make the panning feel natural. The movement is coupled to the phone’s accelerometer – the update frequency needs to be high for fluent movements. At the same time, it needs to ignore small trembling movements.
Context about the project:
Neue Zürcher Zeitung is a Swiss daily newspaper with a focus on quality journalism and a liberal-conservative readership. Many of our readers are especially skeptical about the transition from nuclear to solar energy.
Most electricity in Switzerland is produced by hydroelectric power plants. Most of the rest are by nuclear energy. In 2016 the “energy strategy 2050” was passed by parliament which bans the construction of new nuclear power plants. One year later the law was also accepted by 58 percent of the population in a direct vote.
The article is based on a revised version of the predictions from 2021 which addresses major critiques of the original scenario. Specifically, it includes predictions on the growth of e-mobility and population growth.
Switzerland is not an EU country and has 2021 failed to ratify a set of agreements that would have specified many aspects of its relation to the EU. One part of these agreements touched on energy markets and imports. In the absence of such regulation, it might become difficult for Swiss energy providers to import electricity after 2025. This development made energy independency a more urgent concern recently.
Switzerland is a country of renters. Only about 40 percent of the population live in an apartment/house they own. The other 60 percent are not free to install solar panels for their personal consumption.
What can other journalists learn from this project?
In this project we have used an approach we call *re-specification*. It can be used for any generalization and is based on an observation similar to one Daniel Kahneman notes in his book “Thinking fast and slow”: it is easy for people to generalize from a concrete example. But it doesn’t work the other way around. This principle is especially interesting for data journalists who often need to communicate abstractions.
Even though we believe in the power of this approach, it is very time-consuming and requires a lot of data for context – which would not be needed when simply showing the statistic.
Even though most of our testers were delighted by the novelty of the interaction. They quickly noted that in many places it might be socially unacceptable: Lifting your smartphone and moving it around while sitting on a train immediately makes people think you are filming them. This might prevent many readers from interacting with the panorama leisurely.