📈 USD/IDR Exchange Rate Forecasting Using FBProphet

🥇 Gold Medal — World Youth Invention and Innovation Award (WYIIA 2023)

Mathematics Category | Yogyakarta, Indonesia | October 2023

---

Overview

This research forecasts the Indonesian Rupiah (IDR) exchange rate against the US Dollar (USD) for the next 76 days (September 15 – December 29, 2023) using the FBProphet time-series forecasting algorithm developed by Meta.

Metric	Result
Training data	166 business days (Jan 2 – Sep 14, 2023)
Forecast period	76 days (Sep 15 – Dec 29, 2023)
MAPE	0.006506206%
Accuracy	99.993%
MAPE category	Very Accurate (< 10%)
Trend direction	↑ Upward (IDR weakening against USD)

---

🥇 Award

This project was awarded the Gold Medal at the World Youth Invention and Innovation Award (WYIIA) 2023, organized by the Indonesian Young Scientist Association (IYSA) in collaboration with Universitas Sarjanawiyata Tamansiswa (UST) and Institut Pertanian Bogor (IPB).

📄 Full extended abstract: Available upon request

---

Research Team

Name	Institution
Arditya Sulistya Ningsih Apusing (Lead)	Tadulako University
Rifdah Apriliani	Tadulako University
Zhafirah Tri Nursagita	Tadulako University
Uswatun Hasanah	Tadulako University
Muhamad Aidil	Tadulako University

---

FBProphet Model

FBProphet uses an additive decomposition model:

y(t) = g(t) + s(t) + h(t) + ε(t)

Component	Description
g(t)	Trend — non-periodic changes over time
s(t)	Seasonality — weekly/annual periodic patterns
h(t)	Holiday effects — irregular schedule events
ε(t)	Error term — unusual changes not captured by model

---

Key Findings

• IDR/USD shows a downward trend Jan–Sep 2023 (rupiah strengthening), then reverses upward in the 76-day forecast period
• Wednesday consistently shows the highest IDR weakening (weekly component)
• Forecast predicts continued IDR weakening through December 2023
• Model achieves 99.993% accuracy — MAPE categorized as "Very Accurate" (< 10%)

---

Repository Structure

usd-idr-forecasting-fbprophet/
├── WYIIA_USD_IDR_Forecasting_FBProphet.ipynb   ← main analysis notebook
├── data/
│   └── Data_Kami_WYIIA.xlsx                    ← USD/IDR dataset (Bank Indonesia)
├── paper/
│   └── WYIIA2023_Extended_Abstract.pdf         ← published extended abstract
└── README.md

---

How to Run

# 1. Install dependencies
pip install prophet pandas matplotlib openpyxl jupyter

# 2. Clone repository
git clone https://github.com/ardityaapusing/usd-idr-forecasting-fbprophet.git
cd usd-idr-forecasting-fbprophet

# 3. Open notebook
jupyter notebook WYIIA_USD_IDR_Forecasting_FBProphet.ipynb

---

Data Source

Bank Indonesia Official Website — bi.go.id

• Variable: Daily USD/IDR middle exchange rate (Kurs Tengah Bank Indonesia)
• Period: January 2, 2023 – September 14, 2023
• Total: 166 business days
• Data is publicly available at: https://www.bi.go.id/id/statistik/informasi-kurs/transaksi-bi/default.aspx

---

Descriptive Statistics (Training Data)

Statistic	Value
Mean	15,102.6205 IDR/USD
Variance	49,378.8
Maximum	15,635 IDR/USD (Jan 6, 2023)
Minimum	14,632 IDR/USD (May 4, 2023)
Standard Deviation	222.213344

---

Tools & Stack

Python 3.11 · Prophet (FBProphet) · Pandas · Matplotlib · Jupyter Notebook · openpyxl

---

Acknowledgments

• Indonesian Young Scientist Association (IYSA) — competition organizer
• Bank Indonesia — data source
• Prof. Junaidi, M.Si., Ph.D. — research supervisor
• Ade Hendrawan Krisdianto, S.Stat. — methodology guidance

---

Citation

Apusing, A.S.N., Apriliani, R., Nursagita, Z.T., Hasanah, U., & Aidil, M. (2023).
Forecasting The Dynamics Of The Indonesian Rupiah (IDR) Exchange Rate Against
The Dollar (USD) Using FBProphet Algorithm Analysis.
Extended Abstract, World Youth Invention and Innovation Award (WYIIA) 2023.
Gold Medal, Mathematics Category. Yogyakarta, Indonesia.

---

Faculty of Natural Science and Mathematics, Tadulako University, Palu, Indonesia Contact: ardityasulistya6@gmail.com | linkedin.com/in/ardityaapusing