A dual-engine Agricultural Intelligence Platform designed to forecast rice productivity across five provinces in Kalimantan and detect crop diseases. The system combines climate data from NASA POWER and agricultural statistics from BPS (Badan Pusat Statistik) for forecasting, alongside image datasets sourced from Kaggle for deep learning-based plant health diagnostics.

• The problem: Rice farming in Kalimantan faces two major threats: high sensitivity to climate variability and the devastating impact of crop diseases. Planners often rely on lagging historical statistics, making proactive strategy difficult, while farmers lack accessible tools for early disease identification.
• Our approach: We implemented a dual-model architecture. For forecasting, we developed a multivariate LSTM model as our primary predictor, benchmarked against Linear Regression, Random Forest, and XGBoost. For disease detection, we utilize a Convolutional Neural Network (CNN) trained on Kaggle datasets to accurately classify leaf imagery. Both models are served through a FastAPI backend and seamlessly integrated into an interactive React/Vite web dashboard.
• Who it's for: Agricultural planners, local government stakeholders, and researchers who need actionable productivity forecasts, as well as field agronomists and farmers requiring quick, data-driven plant health diagnostics.

No	ID	Member	Role	Connect
1	APC524D6Y0192	Muhammad Rahman	Machine Learning Deployment
2	APC747D6X0300	Siti Fitria Putri Hans	AI Engineer
3	APC216D6Y0441	Muhammad Rakha Djauhari	Project & Quality Assurance
4	APC480D6Y0457	M. Rohid Rivaldi	Front-end AI Integration
5	APC524D6Y0481	Muhammad Husain Ali Ridha	AI Engineer

Special thanks to the parties that supported our capstone journey:

• Pijak Program — Independent study framework & program structure
• IBM SkillsBuild — AI Engineer curriculum & learning resources
• Dicoding — Program operations & learning platform
• Our Facilitators — Guidance during ILT sessions and continuous mentoring
• Our Mentor Advisors — Project insights, direction, and constructive feedback
• Our Academic Advisors — Administrative support from our respective universities

This project is licensed under the GNU Affero General Public License v3.0 (AGPLv3) — see the LICENSE file for full details.

While we welcome open collaboration through forks and pull requests—and encourage you to reach out if you plan to build something significant upon this project—please note that our model predictions are strictly experimental and must not serve as the sole basis for real-world agricultural or policy decisions.