Data model for legal and ethical compliance by design in data science projects

Resumen

Data is the most important ingredient in any data science project. It is also a fundamental aspect of any software development project. Considering the required data specifications the assessment of legal compliance is critical. It is also necessary to evaluate the social and ethical impact of data to avoid undesirable results and consequences. This project aims to provide sufficient vocabulary in the form of an ontology that allows data publishers to describe their data in terms of legal, social, and ethical impact. Data publishers should be able to explain the main purpose and the legal basis for collecting the data, the risks associated with data processing, and the ethical assessments that have been performed. With all this information, data scientists, software developers and other interested parties can decide whether the data is suitable for their projects. The main approach to create the ontology was to follow the Linked Open Terms Methodology. This ontology engineering methodology divides the process into four parts: ontology requirements specification, ontology implementation, ontology publication, and ontology maintenance. After completing all the necessary steps, an ontology has been created and published on https://protect.oeg.fi.upm.es/def/lseid. With this vocabulary data publishers are able to provide all the relevant information about legal, social, and ethical aspects of the data. This thesis describes in detail the whole process of creating the ontology. It also explains design choices and presents the final result.