Prompt

The state of social media research APIs & tools in the DSA era

The state of social media research APIs & tools in the DSA era

19/12/2024

social-networks

Our white paper analyzes how the European Union's Digital Services Act (DSA) is transforming data access for platform research, drawing on existing literature and extensive testing of research APIs and tools from major platforms including Meta, YouTube, and TikTok. Our insights are based on years of experience working with social media data and addressing critical data quality issues in platform-provided datasets.

We focus particularly on Article 40.12, which mandates access to "publicly available data," examining how Very Large Online Platforms (VLOPs) are implementing these requirements. The analysis identifies key gaps, challenges, and potential solutions, emphasizing recent developments not yet covered in existing reports.

Key highlights include:

  • Critical advancements and challenges in DSA data access, addressing emerging issues and opportunities not yet extensively explored;

  • Detailed evaluation of VLOPs' compliance with Article 40.12, including platform-specific analyses of APIs, tools, and data-sharing mechanisms;

  • Implementation challenges, including technical constraints, restrictive terms of use, and ambiguities in compliance frameworks;

  • Actionable improvements for standardizing data formats, increasing transparency, and expanding researcher accessibility. These recommendations aim to enhance data access usability while fostering a robust research ecosystem.

The white paper concludes by emphasizing the importance of building a strong researcher community to test, evaluate, and improve data access tools. This community will be vital for ensuring tool reliability, advocating for improvements, and fostering collaboration between academia, civil society, and regulatory bodies - ultimately helping realize the DSA's vision of a more transparent and accountable digital ecosystem.

→ Download the white paper

Authors: Fabio Giglietto and Massimo Terenzi, University of Urbino Carlo Bo. This paper was edited by Anwesha Chakraborty (vera.ai) and Martin Lestra (opsci.ai).