Implementación de un repositorio y geovisor de datos meteorológicos e hidrológicos con predicción en la sierra centro.

This research project developed a repository and a geovisor (geospatial visualization tool) with predictive capabilities for meteorological and hydrological data in the central highlands of Ecuador, through a web platform integrated with spatial databases and climate models, focusing on the provinces of Tungurahua, Bolívar, and Chimborazo. The methodology combined the analysis of Geographic Information Systems (GIS) tools with the training of climate prediction models such as Linear Regression, Random Forest, LSTM, and CNN, integrating the use of PostgreSQL and GeoServer. The functionality evaluation of the geovisor, based on the ISO/IEC 25010 standard, showed that completeness and functional correctness placed 58% of responses in the “completely satisfied” category and 42% in “very satisfied.” As for functional suitability, 83% of respondents were “completely satisfied” and 17% “very satisfied.” Among the models evaluated, the Convolutional Neural Network (CNN) outperformed the others in all performance metrics, achieving a coefficient of determination (R²) of 0.89, MAE of 0.82, MSE of 1.17, RMSE of 1.08, and EVS of 0.89, positioning it as the most accurate model with the lowest error rate for maximum temperature estimation.

• 1. Verfasser: Malan Obando, Edison Adrián (author)
• Format: bachelorThesis
• Sprache: spa
• Veröffentlicht: 2025
• Schlagworte: Geovisor; Predicción climática; Repositorio de datos; Sistemas de información geográfica
• Online Zugang: http://dspace.unach.edu.ec/handle/51000/15620