University of Zambia Electronic Theses and Dissertations Dataset: Metadata, Full-Text Extracts, and Ground Truth Annotations

University of Zambia Electronic Theses and Dissertations (UNZA ETD) Dataset Metadata, Full-Text Extracts, and Ground Truth Annotations for 5,725 ETDs Overview This dataset contains 5,725 Electronic Theses and Dissertations (ETDs) harvested from the University of Zambia Institutional Repository (RIMS). It comprises three interrelated components: OAI-PMH Metadata: Dublin Core metadata harvested via OAI-PMH (5,725 records) Full-Text Extracts: Extracted sections from merged PDFs including title pages, abstracts, acknowledgements, table of contents, dedication, and lists (5,208 records) Ground Truth Annotations: Manually transcribed metadata for 3,268 ETDs, used for validation of automatic extraction methods Key Features Comprehensive coverage: 5,725 ETDs from the University of Zambia ETD-ms compliant: Fields follow the NDLTD ETD-ms metadata standard Rich ground truth: 3,268 manually transcribed records (inter-annotator agreement >0.85 Cohen's κ) Augmented metadata: Dublin Core fields enriched with automatic and manual sources Interoperable format: All CSV files use pipe (|) separators Reproducible: Complete Google Colab notebook included for full pipeline reproduction File Structure File Name Records Description db-datalab26-unza_etds-oai_pmh_metadata.csv 5,725 Harvested OAI-PMH metadata db-datalab26-unza_etds-oai_ore_extracts.csv 5,208 PDF section extracts (title, abstract, acknowledgements, etc.) db-datalab26-unza_etds-auto_ground_truth.csv 5,208 Rule-based automatic ground truth db-datalab26-unza_etds-ground_truth_manual_cleaned.csv 3,268 Manually transcribed ground truth db-datalab26-unza_etds-merged_dataset.csv 5,725 Merged dataset with augmented Dublin Core and ETD-ms fields db-datalab26-unza_etds-codebook.csv N/A Data dictionary with column descriptions Citation If you use this dataset, please cite: Phiri, L. (2026). University of Zambia Electronic Theses and Dissertations (UNZA ETD) Dataset: Metadata, Full-Text Extracts, and Ground Truth Annotations. [Data set]. Zenodo. doi.org] Related Publications Chileshe, E., & Phiri, L. (2024). Zambia National ETD Portal: Implementation of an Interoperable National Electronic Thesis and Dissertation Portal. ETD 2024.URL: docs.ndltd.org Kasonde, C., & Phiri, L. (2023). Assessing and promoting metadata quality for electronic theses and dissertations in institutional repositories using a policy-driven approach.URL: ir.inflibnet.ac.in Kasonde, C., Chisale, A., & Phiri, L. (2022). Empirical Evaluation of ETD-ms Compliance in Institutional Repositories. 25th International Symposium on Electronic Theses and Dissertations (ETD 2022), Novi Sad, Serbia.DOI: doi.org Habukali, M., Mbewe, M., Mwale, N., Mwewa, M., Sikazindu, N., & Phiri, L. (2021). Effective Ingestion of Digital Objects in Institutional Repositories Using Subject Repositories. 2021 IST-Africa Conference (IST-Africa).URL: ieeexplore.ieee.org Phiri, L. (2020). Automatic classification of digital objects for improved metadata quality of electronic theses and dissertations in institutional repositories. International Journal of Metadata, Semantics and Ontologies, 14(3), 234-248.DOI: doi.org Phiri, L. (2018). Research visibility in the global South: towards increased online visibility of scholarly research output in Zambia. IEEE International Conference in Information and Communication Technologies.URL: rims.unza.ac.zm Chisale, A., & Phiri, L. (2024). Identification of Sources for Missing Electronic Theses and Dissertations Metadata in Higher Education in Zambia. Master's thesis, University of Zambia. Kasonde, C. (2024). A Study on Policy-Driven Strategies for Enhancing Metadata Quality in Electronic Theses and Dissertations. Master's thesis, University of Zambia. Muchinga, M. (2024). Effective Approaches for Improving the Uptake of Institutional Repositories Content in the Higher Education Institutions in Zambia. Master's thesis, University of Zambia. Acknowledgements The author thanks the DataLab Research Group and the University of Zambia Library for providing access to the repository and supporting the manual transcription efforts. License This dataset is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).