The TrenDTF project set out to investigate how well the outcomes of funded research projects align with the political and scientific objectives set out in German federal funding programmes. The original plan was to collect the full texts of calls for proposals from several major German funding agencies and to compare them with policy documents such as the High‑Tech Strategy and progress reports. The comparison would have relied on a similarity analysis of the content, but the lack of reliable metadata linking project reports to their respective calls made this approach infeasible. Incomplete project identifiers in the reports and the absence of comprehensive lists of programme IDs meant that the necessary cross‑referencing could not be performed.

To address the research question on a broader level, the team turned to a text‑based analysis of policy approaches, announcements, and project reports drawn from grey literature. The primary data sources were the calls for research and innovation programmes issued by selected ministries and the final reports of projects funded under the federal programmes. The methodology centred on text mining, specifically probabilistic topic modelling, keyword extraction, and document similarity analysis. The team reviewed related literature on topic modelling and other machine‑learning techniques for scientific and policy data, and they incorporated empirical studies on mission orientation and impact assessment. The results were validated in a case study of logistics‑related research at the Fraunhofer Institute for Material Flow and Logistics (IML), where funding announcements and final research outputs were mapped and analysed.

In the first work package, a comprehensive literature review was conducted. The review covered grey literature on data sources and their evaluation, making the European Tender Electronic Daily (TED) database accessible for the project. It also examined methodological literature on data analysis, particularly topic modelling, and gathered empirical studies on mission orientation and impact measurement. The findings from this review directly informed the methodological work and were documented in the final report. The second work package focused on database construction. Three groups of data were assembled and stored in databases: policy documents and strategies, public tender data from TED, and project reports from the Technical Information Library (TIB) in Hannover. The TED data, comprising a large and complex dataset, were downloaded and migrated into a suitable database, although linking them to specific German funding programmes remained problematic. The TIB data were integrated into the Fraunhofer Institute for Information Systems (ISI) servers to enable topic modelling. By the end of the project, information from the funding catalogue and identified topics from the promotion database were used for validation.

The project’s technical achievements include the successful implementation of unsupervised topic modelling on large, unstructured policy and research documents, the extraction of key themes and trends, and the development of a framework for assessing the alignment between research outputs and policy objectives. Although explicit performance metrics such as topic coherence scores are not reported in the summary, the methodology was validated against a real‑world dataset from Fraunhofer‑IML, demonstrating its applicability to logistics research. The project also produced a set of validated databases that combine policy documents, tender announcements, and project reports, providing a foundation for future analyses of research impact.

Collaboration involved several German research institutions. The Fraunhofer Institute for Material Flow and Logistics (IML) led the validation effort and provided domain expertise in logistics research. The Fraunhofer Institute for Information Systems (ISI) handled the technical implementation of the databases and the topic‑modelling pipeline. The Technical Information Library (TIB) in Hannover supplied the project reports, while the Leibniz Information Centre for Technology and Natural Sciences contributed to the literature review and data curation. The project ran through 2021, with ongoing literature monitoring and database updates. Funding was provided by German federal programmes, although the specific grantor is not named in the excerpt. The project’s outcomes were disseminated through reports, conference contributions, and participation in workshops, ensuring that the findings reached both the scientific community and policy makers.