Análisis estadístico exploratorio e inferencial de series temporales para la elaboración de reportes metereorológicos de la provincia de Chimborazo en un aplicativo web, periodo (2014-2019)

Abstract

The objective of this research work was to do an exploratory and inferential analysis of time series of environmental temperature monitored by the Group of Alternative Energies and Environment (GEAA) of the Escuela Superior Politécnica de Chimborazo (ESPOCH) for the preparation of meteorological reports in a web application with R Shiny. The data of the environmental temperature of the period 2014 - 2019 provided by the GEAA was used, from the exploratory analysis of the data it was possible to show that the Atillo station presents the lowest temperature and the Cumandá station the highest average temperature. In addition, the decomposition of the time series was done and it was observed that there is a presence of seasonality in each time of the year. Through the study of time series, applying the Box Jenkins methodology, the best ARIMA models were found for each meteorological station and for the validation of the models found, then the assumptions of normality, homoscedasticity and independence to errors, as well white noise were verified. Finally, in the interface of the web application created in R Shiny, graphic predictions were made using the models determined for each season and it was possible to appreciate that in the month of February there is an increase in the average temperature and later it tends to drop until the month of August, while from September to November the temperature tends to rise. It is recommended to the GEAA to apply the models determined in this study, to improve the forecasts of the temperature of the meteorological stations.