<![CDATA[Digital transformation in the aquaculture sector, particularly in catfish farming, holds significant potential to improve operational efficiency and farm productivity. This study developed an artificial intelligence (AI)-based monitoring system called NusAIra to assist farmers in managing ponds in real-time by monitoring water quality, feed management, and harvest prediction. The system integrates physical sensors with a Decision Tree Regression machine learning algorithm, validated using an 80:20 hold-out split strategy and evaluated through accuracy and Root Mean Square Error (RMSE) metrics. NusAIra was built using Flask and Docker frameworks, employing a POST endpoint with JSON-formatted data for seamless data exchange. Implementation was carried out on three catfish ponds in Jepara Regency from February to April 2025. The predictive model achieved an accuracy of 87% with an RMSE of 0.35. One application example demonstrated that the system reduced the Feed Conversion Ratio (FCR) from 1.9 to 1.6, increased productivity by up to 22%, and lowered average operational costs by 15%. Additionally, NusAIra effectively predicted market prices with stable seasonal patterns, such as the projected catfish price in Boyolali for April reaching IDR 36,442, closely aligning with historical data. These results highlight NusAIra’s role in supporting data-driven decision-making. However, challenges remain, including infrastructure constraints and the low level of digital literacy among traditional fish farmers. Further development will focus on enhancing prediction accuracy, integrating adaptive features, and expanding system reach through cloud computing to support the sustainability and food security of Indonesia’s aquaculture sector.]]>
