Mapping culture and arts adult education using announcement texts in databases applying natural language processing methods

Abstract Reporting and research in the field of continuing education are characterised to a large extent by a lack of reliable and comprehensive data. This applies in particular to the supply side of continuing education. This article presents a data source that has been used very little to date for empirical research into continuing education and educational reporting. Using the example of culture and arts adult education, this study demonstrates how natural language processing methods can be used to thematically structure, identify and analyse a very large corpus of non-standardised announcement texts in course databases such as IWWB. The choice of adult education in the field of culture and arts education is particularly relevant, as this area is characterised to a large extent by conceptual challenges of definition, significant data gaps and a heterogeneous provider structure. The proposed approach is also relevant to other educational topics, such as political education and democracy education, education for sustainable development, integration, artificial intelligence or health education, to name but a few, where similar challenges exist regarding the availability and comparability of data in adult and continuing education.