A harmonized longitudinal academic records dataset for higher education analytics

Published: 14 August 2026| Version 5 | DOI: 10.17632/66m55bj4zp.5
Contributors:
, Ngo Thanh Doan, Hoang Duy Kien,
,

Description

This repository contains anonymized and harmonized longitudinal academic data derived from undergraduate students enrolled in an automotive engineering program at an application-oriented university in Vietnam. The 1,656 student records were drawn from five admission-year cohorts and multiple curriculum versions. Differences in course naming, coding, semester placement, and elective pathways required data cleaning, standardization, and cross-cohort harmonization across curriculum versions. Two parallel final datasets are provided, each containing 1,656 student records and 56 variables: a labeled version for direct interpretation and a harmonized coded version for analytical reuse and reproducible workflows. The released variables include student-level contextual information, 50 harmonized course-grade variables, cumulative academic performance, scholarship-related information, and academic-status variables. The two 56-variable datasets do not contain a separate admission-year or cohort identifier. The repository also includes course and variable dictionaries, an executable Python preprocessing file, a synthetic source-formatted course-grade dataset, aggregate audit summaries, a de-identified source-course extract documenting the pre-harmonization data structure, a README file, and supplementary semester-based and subject-group analytical outputs. The public reproducibility materials illustrate the programmatic course-grade harmonization stage only. They do not reproduce the private database extraction, administrative-record merge, manual reconciliation, anonymization, or final 56-variable coding steps. The de-identified source-course extract contains generalized metadata. Exact cohort and admission-year labels, group sizes, and links between the original and public groupings are not released. The original institutional administrative records and institution-specific harmonization mappings are not publicly available.

Files

Institutions

Categories

Education, Data Mining, Machine Learning, Data Analytics

Licence