Süddeutsche Zeitung is the leading daily quality newspaper in Germany. Founded 1945 and based in Munich, it offers reporting, analysis and opinion for a national audience. Editorial staff is 400+ people, with bureaus in Berlin, Frankfurt, Düsseldorf, Leipzig, Hamburg and Stuttgart, plus correspondents in major capitals all around the world. Its award-winning Data Team was established in 2018 to focus on data-driven reporting across all topics, with close links to the Investigations team (known for the Panama Papers, among other stories). Although it is a standalone unit of the newsroom, the team always works very closely with specialist editors from the other departments, graphic designers and developers, combining their expertise and skills with their own to achieve the best result for readers. As the team sees it, data journalism is first and foremost journalism – its goal is to tell stories, explain complex issues, expose injustice and corruption. The team is committed to constantly learning, experimenting with new tools, sources and storytelling formats, and providing the best possible experience for Süddeutsche Zeitung readers – online and in print. Exchanging ideas, sharing knowledge and learning from each other are important pillars of the philosophy. Towards their readers, they try to be as transparent as possible – publishing detailed descriptions of their methodology, source code and raw data wherever possible.
Description of portfolio:
How content is positioned in social media determines how many people see it. Facebook and Instagram have long since stopped displaying their content simply according to the rule “the latest post appears at the top”. Instead, they use their own opaque algorithms that are based on criteria such as the user’s behavior and preferences. But to what extent does this influence how they perceive the world? What kind of content is shown to users, and what is not?
In the #wahlfilter project, the SZ tried to gain insight into how social media works and understand how it influences us – especially before the federal elections in Germany in september 2021. For this, SZ collaborated with the American research platform The Markup as well as Algorithmwatch to collect data from a panel of around 400 users who have been deliberately put together in a demographically diverse way, and show what content is displayed in each person’s Facebook or Instagram feed. This allowed the SZ team to check, for example, how often supporters of political parties see posts from their own favorite party – including its local chapters, or how visible posts from political parties are. Since large parts of the election campaign takes place on social networks, and many people use Facebook and Co. as their main source of information, the question of the influence of their algorithms has enormous social and democratic relevance. And the analyses show: There are differences.