Implementation of an Architecture for Streaming Data Acquisition Incorporating Quality Attributes

Abstract

The volume of data that humanity is producing is a phenomenon that arouses research interest, considering different aspects and contexts involved. In this subject, data quality requires attention for acquisition and proper handling in order to provide accurate results in analytical processes. Nevertheless, data from sensors are sensitive to intercurrences that can compromise decisions guided by indicators constructed in this process. The objective of this work is to implement an architecture capable of producing a layer of data quality associated with the data collected from sensors, in real time. The proposed architecture evaluates the dimensions of accuracy, reliability, completeness, data volume and punctuality. The implemented artifacts were adapted for testing with indoor temperature and humidity measurement sensors. The collected data received quality indicators calculated in the sensor controller, packaged by time windows and transmitted directly to the server. Preliminary results showed efficiency in the acquisition and generation of quality indicators. The final product can be used to other contexts that use data acquisition through sensors.