PoliTok-DE: A Multimodal Dataset of Political TikToks and Deletions From Germany

2025-09-19

AI Summary

This study addresses two critical gaps: insufficient transparency in platform content governance and the lack of multimodal analysis in political communication. To bridge them, we introduce PoliTok-DE—the first multimodal political TikTok dataset focused on the 2024 Saxony state election in Germany. Leveraging the TikTok Research API and compliant web crawling, we collected 195,000 video posts, each with synchronized audio, visual, textual, and metadata components; notably, 17.3% (≈18,000) were subsequently removed by the platform—the first systematic capture and public release of deleted multimodal political content. The dataset ensures reproducible access via post ID lists and data-retrieval code. Through illustrative case studies, it reveals tensions between algorithmic intolerance and political entertainment. PoliTok-DE thus provides a foundational resource for computational social science research on platform governance, political communication, and multimodal methodology.

Technology Category

• Machine Learning: Multimodal Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data

Application Category

• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Security and Privacy: Data transparency and provenance
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

We present PoliTok-DE, a large-scale multimodal dataset (video, audio, images, text) of TikTok posts related to the 2024 Saxony state election in Germany. The corpus contains over 195,000 posts published between 01.07.2024 and 30.11.2024, of which over 18,000 (17.3%) were subsequently deleted from the platform. Posts were identified via the TikTok research API and complemented with web scraping to retrieve full multimodal media and metadata. PoliTok-DE supports computational social science across substantive and methodological agendas: substantive work on intolerance and political communication; methodological work on platform policies around deleted content and qualitative-quantitative multimodal research. To illustrate one possible analysis, we report a case study on the co-occurrence of intolerance and entertainment using an annotated subset. The dataset of post IDs is publicly available on Hugging Face, and full content can be hydrated with our provided code. Access to the deleted content is restricted, and can be requested for research purposes.

Problem

• Analyzing deleted political TikTok content in Germany
• Studying intolerance and political communication on social media
• Developing methods for multimodal research on deleted posts

Innovation

• Multimodal dataset from TikTok API
• Web scraping for deleted content retrieval
• Annotated subset for intolerance analysis