Hotspot clustering using DBSCAN algorithm and shiny web framework

Forest fires are a serious problem that occurs repeatedly in Indonesia. Fire events can be predicted by monitoring the datasets of hotspots which are recorded through remote sensing satellite. This study aims to build a web application that performs clustering on the hotspots data. This application implements DBSCAN algorithm using Shiny web framework for R programming language. Clustering is performed on a dataset of hotspots on Kalimantan Island and South Sumatra Province in 2002-2003. The spread pattern of hotspots resulted by this clustering can be used as a predictive model of forest fires occurence and can be accessed through the internet browser.