Uma proposta para análise de dados abertos do PROCON utilizando Data Mart

Abstract

Nowadays, it has been generated a large amount of data. There are many government systems that provide these data public. However, most of these data available are not legible, and they are incoherent and inconsistent. Thus, it is necessary to propose solutions that manipulate these data, make them understandable as information, and provide them to the public community. This study used data from the Fundação de Proteção e Defesa do Consumidor (PROCON), available to the public on Portal de Dados Abertos do Governo Federal. To transform this data into useful information, we performed all the process of Knowledge Discovery in Databases (KDD) and we proposed a Data Mart using ROLAP (Relational Online Analytical Processing) as a model for the construction. The tests through OLAP (OnLine Analytical Processing) queries were processed on the Data Mart proposed. The results of these queries enabled the generation of graphs that facilitate the understanding of the information. Still on the data analysis, the authors attempted to create a recommendation system based on user profiles of PROCON. However, it has not been possible to establish the profile of users using such open data. As future work, we plan to collect new data for more time analysis, and verify the possibility of create a recommendation system with the integration of other databases related to PROCON. Finally, this work was presented as short course on the Workshop de Tecnologia da Informação do Sertão Central (WTISC 2014) that occurred on Federal University of Ceará, Campus Quixadá.