DMP : Predicting the Legal Basis of Internet Network Locks in Austria Using Machine Learning

Context and methodology

Research Domain: This dataset and DMP were created within the context of the Data Stewardship (DaSt) 2026 course at TU Wien.

Purpose: The primary goal is to investigate if the legal basis for internet network locks in Austria—specifically distinguishing between copyright infringements and EU sanctions—can be accurately predicted using machine learning.

Creation Methodology: The project utilizes an existing regulatory dataset from the Austrian RTR (Network Locks register). I enriched the raw CSV by extracting features such as Top-Level Domains (TLD) and decision years, then implemented a classification pipeline using Orange Data Mining.

Technical details

Structure: The deposit consists of this Data Management Plan (DMP) in PDF format. The associated experiment data is organized into raw data, an enriched CSV for training, and output folders for visualizations and model files.

Naming Convention: Files follow the specific course requirements, using the format <student_id>-use-case-description.pdf and meaningful labels for all exported images and models.

Software Requirements: To interact with the workflow file (.ows) or the trained model (.pkcls), Orange Data Mining software is required.

Additional Resources: Detailed documentation regarding the logic of the experiment and the data flow is provided within the "Use Case Report" and the visual workflow nodes.

Further details

Reusability: This dataset contains no personal information and is based on public regulatory records. It is designed to be a reproducible example of a FAIR-compliant machine learning pipeline. Users should note that the actual model performance is secondary to the demonstration of proper data management practices.