Replication data and code for: Explainable Efficiency-Diagnosis (EED) pipeline for DEA efficiency of Chinese higher vocational colleges

Reproducible analysis code, a complete data dictionary, and an illustrative (synthetic) 35-college sample dataset for the Explainable Efficiency-Diagnosis (EED) pipeline (GBDT + SHAP attribution of DEA technical efficiency). The empirical results in the companion manuscript are computed on licensed institutional data that cannot be redistributed; the bundled CSV shares the same schema so the code is fully runnable and the method transparent. The national reference set is the 2024 cross-section of 1,562 Chinese higher vocational colleges from the national institutional data platform (https://hve.emic.edu.cn).