The larger the number of attributes and legal documents to be analyzed, the more difficult it is for a human to identify interesting relationships between the distinguished attributes and documents that are specific or in some way different from other legal documents. The research team aims to create a recommendation tool that will help to reveal relationships between legal sources and select documents worthy of further legal analysis. This may lead to the identification of so-called outliers, i.e. legal documents (e.g. court decisions) that deviate from established decision-making practice, or enable the discovery and verification of legal hypotheses reflecting internal relationships between legal documents. Case studies focusing on selected legal problems (in the field of personal data protection and criminal law) will be prepared for the purpose of verification.