Perbandingan Pengembangan Model Time Series Forecasting di Lingkungan Lokal (Vscode) dan Cloud Low-code (Aws Sagemaker Canvas)

Hadriansyah, Muhammad (2025) Perbandingan Pengembangan Model Time Series Forecasting di Lingkungan Lokal (Vscode) dan Cloud Low-code (Aws Sagemaker Canvas). Project Report. [s.n.], [s.l.]. (Unpublished)

Text 5025221239-Project_Report.pdf - Accepted Version Restricted to Repository staff only Download (701kB) | Request a copy

Abstract

Kerja Praktik ini dilaksanakan di PT Telekomunikasi Selular (Telkomsel) pada divisi Data Solutions and Digital Financial Services. Kebutuhan akan prediksi traffic jaringan yang akurat menjadi krusial. Proyek ini berfokus pada pengembangan model Time Series Forecasting untuk memprediksi volume penggunaan data atau traffic_gb pada 2000 tower Telkomsel. Metodologi yang digunakan mencakup dua pendekatan komparatif: pertama, pengembangan model scara manual di environment lokal, dan kedua, pengembangan menggunakan platform low-code yaitu Amazon SageMaker Canvas. Kedua pendekatan ini dibandingkan, di mana hasil model lokal unggul dalam akurasi skor metrik RMSE 24,5, sementara Amazon SageMaker Canvas unggul dalam kecepatan pengembangan dan kemudahan deployment, walaupin akurasi skor metrik cukup tinggi yaitu RMSE 145,8. Akhirnya, proses deployment dari SageMaker Canvas didokumentasikan sebagai studio kasus alur kerja cloud low-code.
===================================================================================================================================
This internship was conducted at PT Telekomunikasi Selular (Telkomsel) in the Data Solutions and Digital Financial Services division. The need for accurate network traffic predictions is crucial. This project focused on developing a Time Series Forecasting model to predict data usage volume, or traffic_gb, on 2,000 Telkomsel towers. The methodology used included two comparative approaches: first, manual model development in an on-premises environment, and second, development using the low-code platform Amazon SageMaker Canvas. The two approaches were compared, with the on-premises model yielding an accuracy metric score of 24.5, while Amazon SageMaker Canvas excelled in development speed and ease of deployment, despite a relatively high accuracy metric score of 145.8. Finally, the SageMaker Canvas deployment process was documented as a low-code cloud workflow case study.

Item Type:	Monograph (Project Report)
Uncontrolled Keywords:	Time Series Forecasting, Machine Learning, Cloud Computing, AWS SageMaker Canvas.
Subjects:	T Technology > T Technology (General) > T385 Visualization--Technique T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Muhammad Danis Hadriansyah
Date Deposited:	19 Dec 2025 06:29
Last Modified:	19 Dec 2025 06:29
URI:	http://repository.its.ac.id/id/eprint/129083

Actions (login required)

View Item