ETL Implementation for Centralizing Academic Data in an Educational Bootcamp

Abstract

Inefficiencies in decentralized data management have become a critical issue for bootcamp companies. This study explores the implementation of an Extract, Transform, Load (ETL) process to centralize academic data at PT. Hacktivate Teknologi Indonesia (Hacktiv8) in order to enhance learning process efficiency. With the rapid growth of disorganized academic data, an integrated automation system is required to monitor and evaluate bootcamp participants' progress. The ETL process employed in this study collects data from various Google Spreadsheets, transforms it into a normalized structure, and stores it in a centralized data warehouse using Google Cloud Platform. The research follows the 4-Steps-Kimball methodology to design the data warehouse schema and evaluates data quality based on the ISO/IEC 25012 standard. Data quality assessment is conducted through two approaches: (1) technical validation using Great Expectations shows key attribute accuracy of 98.44%, consistency of 97%, and completeness of 100%, despite 5–7% data loss due to the removal of null values in instructor columns to maintain referential integrity; (2) stakeholder feedback (N=10) collected via a Likert scale (1–5) yields an average score of 4/5, with the highest ratings in data consistency and centralized structure. The evaluation results indicate that the implemented ETL system is highly efficient and positively received by users, although suggestions were made for improvements in system documentation and adaptability. This study is expected to contribute significantly to data-driven decision-making in non-formal education environments.