Spatio-temporal Data Analytics with KNIME

Abstract

The Knowledge Discovery from Data (KDD) process is largely applied to extract new insight from data, and a number of visual platforms, such as KNIME or RapidMiner, offer effective general-purpose tools to data analysts, leveraging a visual programming paradigm. Nevertheless, when dealing with spatio-temporal data, these visual KDD platforms offer only limited support to typical data manipulation tasks, thus forcing data analysts to manually implement several steps of the KDD process, significantly hindering productivity and replicability of the results.

In this talk, we will discuss peculiarities and challenges of implementing Knowledge Discovery processes on spatio-temporal data, drawing from industrial and academic experience. We will then present a solution we are currently working on to support the definition of KDD processes on spatio-temporal data. The proposed solution, which is open source and freely available, extends the KNIME Platform with a collection of new components designed to support spatio-temporal data analytics.