Pembangunan Sistem Otomatisasi ETL Berbasis Web untuk Mendukung Monthly Report

Abstract

Proses penyusunan monthly report di Departemen Asset Accounting PT RS terhambat oleh transformasi data manual dan ketiadaan penyimpanan terpusat yang menyebabkan data historis tersebar, sulit dilacak, dan rentan terhadap human error. Penelitian ini bertujuan untuk membangun sistem otomatisasi Extract, Transform, Load berbasis web yang mengintegrasikan fitur data cleansing, schema validation, centralized repository, dan dashboard pelaporan sebagai pengganti tools Business Intelligence eksternal. Metodologi yang digunakan adalah Rapid Application Development dengan Python (Flask) dan MySQL. Sistem diuji melalui Alpha Testing dengan Black Box Testing teknik Equivalence Partitioning sebanyak 34 test case menghasilkan persentase validasi 100%, Beta Testing kepada 8 responden menghasilkan kelayakan 96,4% (Admin) dan 99,1% (User) berkategori "Sangat Layak", serta Performance Testing dengan error rate 0,00% pada beban normal dan 0,30% pada beban tinggi, serta throughput stabil 3,1 transaksi per menit. Hasil pengujian ini menegaskan bahwa sistem memenuhi seluruh kriteria penerimaan dan layak untuk mendukung proses monthly report yang lebih efisien.

The process of preparing monthly reports in the Asset Accounting Department of PT RS is hampered by manual data transformation and the lack of centralized storage which causes historical data to be scattered, difficult to track, and vulnerable to human error. This study aims to build a web-based Extract, Transform, Load automation system that integrates data cleansing, schema validation, centralized repository, and reporting dashboard features as a replacement for external Business Intelligence tools. The methodology used is Rapid Application Development with Python (Flask) and MySQL. The system was tested through Alpha Testing with Black Box Testing using the Equivalence Partitioning technique for 34 test cases resulting in a validation percentage of 100%. Beta Testing to 8 respondents resulted in a feasibility of 96,4% (Admin) and 99,1% (User) in the "Very Feasible" category, as well as Performance Testing with an error rate of 0,00% at normal load and 0,30% at high load, as well as a stable throughput of 3,1 transactions per minute. These test results confirm that the system meets all acceptance criteria and is worthy of supporting a more efficient monthly report process.