- What is DataSkop?
- Pilot project
- Research and Journalism: Use DataSkop for your data donation project
What is DataSkop?
DataSkop is a platform for data donations. Data donations can help shed light on the black boxes of social media algorithms and automated decision systems. This is demonstrated by examples such as the Citizen Browser —created by The Markup, a US non-profit journalism outfit—which takes a close look at how Facebook works, or the data donation infrastructure Mozilla Rally [link]. The OpenSchufa (2018) and Monitoring Instagram (2020-2021) projects created by AlgorithmWatch—one of the initiators of DataSkop—also use data donations to explore algorithms.
DataSkop provides a technical framework that simplifies many of the repetitive steps required in a data donation project so that the data donation method can be deployed with minimal effort and as cost-effectively as possible. For now, DataSkop runs on desktop and laptop computers, but not on smartphones and tablets.
DataSkop not only allows users to donate data to projects they want to support, but they can also use DataSkop to explore and analyze the data they want to donate themselves. This gives them a better understanding of exactly what they are helping with when they donate data. They also learn what traces they leave behind when they use digital offerings and what makes the data generated in the process so valuable and interesting.
Starting in the fall of 2021, DataSkop will offer learning scenarios for in-school and out-of-school educational settings in addition to the actual data donation. In addition, DataSkop is trying out how the data it collects can be reused for further research while maintaining appropriate privacy protections in the spirit of a data trust.
From 2022, the open-source platform DataSkop will also be open to other research institutions, civil society organizations, and editorial offices, which can present their respective data donation projects there and acquire users willing to donate. With each new project, the DataSkop community grows.
Stay up to date with the Newsletter of AlgorithmWatch, which is going to features news on DataSkop now and then.
“Election recommendations: What does the YouTube algorithm show you about the German federal election?” — is the title of the first project running on the data donation platform DataSkop. It examines which personalized recommendations and search results the YouTube system delivers around topics related to the German federal election. For six weeks, from July 15 to August 25, 2021, data can be donated for this purpose. Unfortunately is only available in German.
What is special about this investigation of the YouTube recommendation system? For one thing, it involves a large-scale analysis of personalization through the video platform’s recommendations system. So far, most research on this has only looked at YouTube recommendations and search results from the outside, i.e., without being logged in with a personal account. On the other hand, potential donors are first shown evaluations and explanations of the data in their YouTube profile. In addition, they can use interactive experiments thanks to visualizations to search for patterns and anomalies in their data. Finally, if they are convinced that they want to use their data to shed light on YouTube’s recommendation system, they can donate their data.
The donated data will then be analyzed by both scientists from the European University Viadrina (Frankfurt Oder) and data journalists from Der Spiegel. The first findings from the data are to be published even before the federal election at the end of September 2021.
A workspace module on the YouTube recommendation system—designed for use in media education in and out-of-school—is also expected to be made available in November 2021.
Research and Journalism: Use DataSkop for your data donation project
If you have an idea for a project, please get in touch with us now. The DataSkop data donation platform is expected to be fully usable and free of charge for third parties from spring 2022. However, projects can also be prepared before that date. Important to know: For the foreseeable future, the platform will only work on laptop and desktop computers and not on mobile devices (smartphones, tablets).
The platform can only be used for non-commercial scientific research projects, civil society campaigns, and journalistic reporting. DataSkop will present a catalog of criteria in the winter of 2021/22 and develop a procedure that transparently regulates access to the platform. Please note: Those responsible for a data donation project, for example, senior employees of a university institute, are fully responsible for complying with data protection and ensuring informational self-determination.
DataSkop significantly reduces the technical and financial effort required to implement a data donation project. There are no fees or similar for the use of the platform itself. Nevertheless, costs are to be expected: In addition to legal advice on data protection issues, for example, software elements have to be developed for the projects, so-called scrapers, which read out data relevant to the respective research question, e.g., from a sales platform. In addition, the development of special visualizations with which users can explore their data before donating it incurs further costs. The respective project must provide the corresponding personnel for the software implementation of the project itself. DataSkop does not offer any corresponding software development.
DataSkop will provide documentation for technical questions regarding the implementation of data donation projects by third parties in the English language in spring 2022. All DataSkop software will be released under an Open Source license. This will include scraper building blocks for specific platforms, such as YouTube, which can be reused or modified for data donation projects by others if necessary. Find the code of the pilot project on Github here.
DataSkop is a joint project, lead and coordinated by AlgorithmWatch. In cooperation with the European New School of Digital Studies (European University Viadrina), we implement our first own data donation projects and determine the functionality of the platform more precisely. The Department of Design of the University of Applied Sciences Potsdam will develop the visualization concepts. The University of Paderborn and the organization Mediale Pfade will develop media pedagogical concepts and workshops in which DataSkop will be used as a learning tool. The three-year project is funded by Germany’s Federal Ministry of Education and Research (BMBF) as part of the program “Mensch-Technik-Interaktion” (human-technology interaction).