The Agricultural Analytics Dashboard Using Data Mining and Machine Learning to Monitor SDG-2 (Zero Hunger) Achievements in Blitar Regency

Abstract The agricultural analytics dashboard is an important platform for monitoring the achievement of SDG-2 (Zero Hunger). The use of data mining and machine learning is the right technique for developing dashboards that are informative and easy to understand by stakeholders. Based on the results of the comparison of harvest area and production prediction models, the Double Exponential Smoothing (DES) method is the best model compared to Autoregressive Integrated Moving Average (ARIMA) and Neural Network (NN). The K-Means method was chosen to group regional potential with Silhouette scores and the Davies-Bouldin Index of 0.40052 and 1.5912 respectively. Based on the results of factor analysis, there are 3 factors that form food security, namely economic growth, food accessibility and food availability. Meanwhile, the best food security level classification model is Randon Forest with Synthetic Minority Over-sampling Technique (RF-SMOTE) with an accuracy value of 0.93 and a precision of 0.95. Use of data mining and machine learning in building agricultural analytics dashboards. The analytical dashboard platform becomes a comprehensive visualization that is easy to understand by agricultural stakeholders to monitor, predict, map and position commodities as a basis for decision making and establishing policies/programs to support national food security.