Crowdsourcing@SZ: Project #MeineMiete/#MyRent

Over the past few years, rents have gone way up in the urban areas in Germany. Moritz Zajonz, who is also part of Journocode, works at the data-driven journalism and storytelling department at the German newspaper Süddeutsche Zeitung. This year, they published an investigation that tried to measure how rent affects people’s lives. Because there was no sufficient data available, they crowdsourced it.

The idea for this project came up, because some of us were personally affected by the rising rents – we are located in Munich where rents are among the highest in Germany. After some internal discussions and talks with experts, we decided that we could, at the very least, attempt to close in on the personal consequences of rent and which circumstances shape people’s decisions. But how? The data that is already openly and not-so-openly available has some shortcomings: The Mietspiegel, roughly translatable to rent index, has no common methodology for two cities, let alone Germany as a whole. Private datasets are available (for a fee), but we didn’t want to rely on data from real estate companies, because of intransparency and conflicts of interest.

In consequence, we went with crowdsourcing, an approach to data generation that had already been on our minds for some time. We spent weeks to generate a questionnaire that readers should answer. To refine it and make sure we had some scientific background for what we were doing, we asked Philipp Doebler, who is a university professor for statistics in Dortmund, for advice. It wasn’t easy to find a balance between our wish for extensive information and the danger of annoying people with too many questions. In the end, we had a questionnaire with about 30 questions, which ranged from multiple choice to free text fields. To be able to differentiate between, for example, usual rent models and flat-sharing communities or sublettings, we build in branches at two points.

When planning the questionnaire’s distribution, we tried to broaden the range of the possible participants, so we wouldn’t exclusively reach readers from So, we placed an illustrated version of the questionnaire in the printed newspaper that readers could fill out and mail back to us. Also, the survey was featured on ad screens in Munich’s subways.

This scatter plot from the Munich longread shows the rent-to-income-ratio of the participants of #MeineMiete. To aid the reader in understanding it, Felix Ebert build it with d3.js, so it could be explained step by step. At the end, the reader could enter a zip code and living status to explore the data. The red line marks the point of rent-to-income-ratio that sociologists describe as precarious.

It was clear though, that we would not be able to draw a representative sample of the German – or Bavarian or any – population. We did this project nonetheless, because we weren’t necessarily looking for hard, absolute numbers, but a glimpse into the lives of more people than just our circle of friends and family. And I think we achieved that: 57 000 people took part in the survey.

The analysis of this massive dataset took quite some time. Foremost, because we had never anticipated such a huge response – especially, the overwhelming amount of personal stories that people shared with us through the free text fields at the end of the questionnaire. This topic was clearly moving people to invest time and energy. So, we did our best to tell as many of their stories as we could. We teamed-up with several more writers to interview people who gave us their consent to be contacted, as well as research and report related stories about the rent situation. A colleague covering the Munich beat, Birgit Kruse, worked with us on a detailed longread about the situation in the city, in which we used the Javascript library d3.js, so readers could find themselves in a scatter plot. And we adapted a tool we had used before so readers could explore the participants’ anonymized stories.

#MeineMiete was a very resource-intensive project. I think, however, that it paid off: We gained experience with crowdsourcing as well as storytelling and we shed a spotlight on this difficult and omnipresent, topic. Maybe most importantly, we were able to give a voice to the many moving stories of tenants who have to jump through hoops just to find a place to live or live in fear of having to leave their home.

This project wouldn't have been possible without Sabrina Ebitsch, Christian Endt, Martina Schories and many more.
You can find the main longread here: and the corresponding stories here: (you need to scroll a bit past recent stories). Some more info on the methodology can be found here: All our texts are in German only, unfortunately.


Moritz Zajonz

Moritz Zajonz works at the German newspaper Süddeutsche Zeitung in the Entwicklungsredaktion. In this data-driven journalism and storytelling department he works on long projects as well as more news-oriented ddj.

Runs on:

How many pie charts have you built?
*An ancient memory pops into his mind. His conscience screams in agony.* “I... - None.”

How many items are on your desktop?
You know, I resisted the urge to just drop stuff on “Desktop” for so long. But I am just a human, weak. So: You can still see some of the wallpaper.

Swear words per day?
Correlates heavily with project intensity and deadlines.

How many adapters do you have?
One to rule them all! Well, and one so I can stick a LAN cable in my other laptop. And there’s also still the problem with VGA monitors...

Your funniest file name?

snow flake
© 2018 Journocode