A BIG DATA ANALYTICS APPROACH FOR FORECASTING AGRICULTURAL COMMODITY PRICES

Abstract

Agricultural commodity prices play a crucial role in economic stability and food security, particularly in developing countries such as Indonesia. Price volatility in key commodities such as rice, chili, and shallots often affects household expenditure, trade balance, and national inflation. Conventional forecasting methods are limited in capturing the complexity and scale of agricultural market data, which is often generated from multiple heterogeneous sources including government reports, wholesale markets, and social media. Big Data Analytics provides an opportunity to address these challenges by integrating large-scale datasets and applying advanced forecasting techniques to generate more accurate predictions. This study proposes a Big Data Analytics framework for forecasting agricultural commodity prices. The framework consists of four main stages: data acquisition from public datasets and online sources, data preprocessing and transformation using distributed computing systems, analytical modeling with machine learning algorithms, and visualization of price forecasts through interactive dashboards. The research implemented Apache Spark for data processing and applied time series forecasting models, including ARIMA and Long Short-Term Memory (LSTM) neural networks, to predict short-term price fluctuations. The experimental results indicate that LSTM outperformed ARIMA in terms of accuracy, with a Mean Absolute Percentage Error (MAPE) of 6.5% compared to 9.8% for ARIMA. Visualization of the forecasts provided clear insights into potential price increases, enabling policymakers, traders, and farmers to make proactive decisions. The novelty of this research lies in the integration of a distributed Big Data processing framework with predictive modeling tailored to agricultural commodity markets in Indonesia. In conclusion, the proposed Big Data Analytics approach demonstrates significant potential to improve forecasting accuracy and support decision-making in agricultural economics. The findings highlight the importance of adopting Big Data-driven solutions for enhancing national food security and market stability.