Sistema de minería de datos del portal web setec para análisis de oferta y demanda al definir indicadores estáticos y dinámicos: sistema de minería de datos del portal web setec para análisis de la demanda al definir indicadores estáticos y dinámicos.

Abstract

The project consisted of developing a data mining system for the National Secretariat of Qualifications and Professional Training (SETEC) web portal to analyze the demand for courses and/or profiles of qualified professionals at a national level. The project was developed using CRISP-DM data mining methodology with the business understanding phase for the analysis of the SETEC web portal modules, understanding of data that make up each module of the web portal, data preparation, data extraction using web scraping techniques, data cleaning, transformation of variable types, and collection of these to create a data warehouse. Subsequently, the demand for these courses and profiles was analyzed based on the definition of static indicators based on real data and dynamic indicators developed through various machine learning models. The result of these phases was displayed in the Microsoft PowerBI tool through various graphics and visual components that show the indicators and their variation with different data dimensions. The usability of the visual application was evaluated based on the System Usability Scale (SUS) and Nielsen's 10 usability heuristics, where 40 people, including experts in data areas and those with moderate knowledge of data analysis and information technologies, scored the application. They concluded that it is an application with very few aesthetic problems and a good usability range.