"predict-dropout-dataset"

"This dataset comprises anonymized academic and demographic records of undergraduate students from the business faculty of a private, highly international university in Turkey, spanning the enrollment years 2011 to 2022. It uniquely integrates institutional student data with external macroeconomic indicators (inflation, unemployment, and political stability) corresponding to each student's home country at the time of their enrollment. The repository contains raw data files (isletme-ogrenciler-en.csv, isletme-ogrenciler-notlar-en.csv) detailing student demographics (e.g., gender, scholarship status, nationality) and semester-by-semester course grades. Additionally, it provides processed datasets (ogr_ml_macros_final.csv) specifically engineered for machine learning applications, where temporal grade data has been pivoted into cumulative GPA features and merged with World Bank economic data. The dataset is designed to support research in educational data mining (EDM) and learning analytics, specifically for developing predictive models of student dropout and retention in international higher education contexts. It serves as the foundational data for the accompanying study on the role of macroeconomic factors in student attrition.Key features:*   Longitudinal Data: Academic records across multiple semesters.*   Macroeconomic Context: Integration of inflation, unemployment, and political stability indices linked to student nationality.*   Machine Learning Ready: Pre-processed CSV files with engineered features suitable for classification tasks.*   Demographic Diversity: Includes data from a diverse international student body.  "