Implementação de técnicas de preparação de dados em índices pluviométricos do Rio Grande do Norte

Abstract

The information generated through applied rainfall measurement has an increasing importance in the scientific, educational and commercial spheres, mainly for helping to develop optimized systems to improve water resources usage. However, rain gauge networks are frequently poor and inadequately distributed, consequently generating historically inconsistent rainfall measurements, obtaining inferior quality than necessary for future applications. Therefore, the execution of rainfall data preparation techniques is necessary to allow later use of historical rainfall measurements. These techniques are designed through the study of data science and, specifically, based on concepts of data preparation. The present work consists of an in-depth analysis of the problem previously described, proposing generic topics focused on processing rainfall data. Subsequently, the initial filtering of raw files from two different sources of rainfall data is detailed, along with a description of Python-developed scripts, which are responsible for preparing rainfall data from different sources and unifying them in filtered and organized form.