Um framework para análise e visualização de dados de proveniência

Abstract

Provenance is recognized today as a central challenge to establish reliability and pro-vide security in computational systems. In scientific workflows, provenance is considered essential to support the reproducibility of experiments, interpretation of results and diagnosis of problems. We consider that these benefits can be used in new contexts, like software process. However, for a better understanding and use, efficient and friendly mechanisms are needed. Software visualization, ontology, and complex networks can help in this process by generating new data insights and strategic information for decision making. This dissertation presents a framework named Visionary, to assist in the understanding and use of provenance data through software visualization techniques, ontologies and analysis of complex networks. The framework captures the provenance data and generates new information using ontologies and analysis of provenance graph. The visualization presents and highlights the inferences and the results obtained with the analysis. Visionary is a context-free framework that can be adapted to any system that uses the PROV provenance model. In order to evaluate the proposal, an experimental study was carried out, which found indications that the framework assists in the understanding and analysis of provenance data, supporting decision making.