The best reason to collaborate is to gain knowledge.

In Tagesspiegel Innovation Lab’s collaborative investigation Cities for Rent, 25 journalists from 16 cities in 16 countries joined forces to shed light on the housing market in European capitals. The goal was to learn about corporate landlords and the financial structures behind the growing market of rental apartments.

Data analysis and data visualisation played a key role in that collaboration. Here’s how we developed a multi-language dataviz environment for all participating media and how it helped structure the investigation.

Bringing the facts together

The initial aim of the investigation was straightforward: We wanted to find out which landlords own the most flats in each of the participating cities. Secondly, we wanted to find out if there are similar actors across different cities, if their tenants had similar problems, and if we could find larger financial structures behind them.

Each editorial team compiled a list of the largest landlords in their city, and we combined that data in a secure, encrypted database. This proved extremely difficult, as not all countries make that information accessible. So combining that data into valuable visualisations across Europe proved difficult — and in some cases impossible.

We found that even the most basic information, like the average rent per city in Europe, is not collected by any governmental institution. It soon became apparent that a map of all the participating cities showing the main corporate landlords per city wasn’t possible.

So we refocused on three aspects:

1. How to show the results of the investigations per city.
2. How to show international connections.
3. How to get proper comparable data.

We developed a visual system that helped solve these problems. 

Journalists provide relevant info

To find out what kind of visualisations are helpful to show the findings for each team in each city, we used conducted simple polls and some in-depths-interviews with  participating journalists.

We asked:

• What content management systems do you use?
• What kind of interactive graphics can be embedded?
• Can you do that yourself?
• Do you have a tech team?
• Do you have style guides?

We also wanted to know what data was the most relevant for each team to show: Was it about developments over time? About spatial distribution across the city? Was it relational data indicating the position of one city in comparison to other cities?

Each team found relevant actors and problematic structures in the housing market of their cities. While the findings varied, they all showed a shared housing problem in capital cities: It gets increasingly financialised, and large international investments into rental apartments have increased since the financial crisis.

Creating a single style 

The technological differences and relevant findings varied so much between teams that a single solution was impossible. To accommodate the different findings and technological conditions, we developed a modular dataviz system that allowed us to embed each graphic individually. Users could easily translate each visualisation from the original English version into the national language of their media organisation.

For these visualisations, we collected all the relevant data compiled by the individual journalists. That quickly showed which information wasn’t comparable and provided some leads for certain cities that occupied a particular position in comparative ranking (e.g., the city with the highest proportion of tenants among the population). The visualisations helped communicate findings to readers and structure some of the research and results within the team.

Our graphics backend also allowed each journalist to adapt the visualisations by highlighting different data points and changing the titles and descriptions. We implemented the different typography of all the organisations, so the visualisations felt like a natural part of their text and even allowed the graphics’ colours to be adjustable.

The results were published by 18 media organisations. Although the visualisations looked slightly different, they had a shared visual language. And while most of the comparative rankings could be used by everyone, many local maps were more specific.

Making it individual through collaboration

One problem in reporting on housing, especially on the financial market behind it, is that it can feel very abstract. Infographics might even reinforce that feeling. After all, it is about the places people live in — their homes.

To overcome that problem, we worked with an illustrator who created a set of illustrations that feel very personal but can serve as teaser images for very different research findings. We developed a colour scheme that was shared across the illustrations and infographics. That way, each media organisation could create visually consistent stories of the different visualisations and illustrations.

On the day of publication, when all the different articles in different languages and countries were released, readers could still see and feel that the research was connected. It was created by one collaborative team, but it benefitted from its members’ regional and cultural diversity.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .