A solution for visualizing decision tree models and association rules in Grafana

e409

Abstract

As a final element of a Data Engineering process, the comprehensible visualization of information is essential. Among the free tools that can be used for visualization is Grafana. Despite its numerous advantages, this tool does not offer graphics that allow the visualization of decision trees or association rules, two of the most popular models for representing the knowledge obtained from a data mining process. This paper presents a solution to this problem that includes, as an intermediate step, the representation of the obtained models in a relational database, as well as the adaptation and extension of Grafana components to display these models. The steps involved are explained, and a case study is provided to evaluate the advantages of the proposal. Consequently, visualizations of each algorithm are obtained with dynamic and customizable elements, which enabled not only the graphical representation of the generated models but also the interpretation of the relationships between the data.