Garantindo a qualidade de dados na fusão de dados semânticos: um caso de uso de SHACL em dados de mobilidade e educação

Abstract

An important factor for the quality of knowledge gained from the data is the amount of data available for analysis and therefore it is ideal that data sets grow and become more useful over time. However, just increasing the number of available data without worrying about quality does not make the data more useful to the user and may even end up rendering the set unusable. When merging two or more semantic datasets, incorrect or incoherent data can be merged, causing the loss of quality of the set as a whole. The dimensions of quality, conciseness, consistency, and accuracy are important to ensure the integrity and veracity of the dataset, and the concept of ontology-based data management (OBDM) can help to ensure that these characteristics are present in datasets that grow over time. Of time. This work aims to create a mechanism that allows these three dimensions of quality to be measured, seeking to provide resources for people who want to increase the number of data to understand whether the data sets used have quality. The mechanism was built in the SHACL language and tested on a set of semantic data from the urban mobility and education domain. A script in Python was also created that allows the execution of this mechanism in a dataset.