Diseño de una arquitectura de datos masivos para el mapeo geográfico de la contaminación del aire en el DMQ entre los años 2005 al 2020.

Abstract

The aim of this research is to integrate the processing and visualization of geographic data related to air pollution, within a massive data architecture. To achieve it, air pollutant data from Quito, Ecuador were used; it is measured by Atmospheric Monitoring Network of the city and were expressed by an index. The proposed architecture is open-source and is made up of a master node and two workers: it consists of a unified analysis computational system in Spark, managed by Yarn and linked to a graphical interface provided by Zeppelin; which stores data in HDFS, processes it and displays it visually through geographic maps. To measure its efficiency, a response time comparison was made with and without the system. Without its use, geographic interpolation took 4.52 seconds; while the proposed system showed an execution time of 2.0 seconds, obtaining a reduction of 56%. This shows an improvement in the traditional interpolation and map visualization processes and generating a new open-source alternative with resources and time optimization; in addition to contribute to making strategic decisions through a new way of analyzing environmental problems.