<p>The rapid changes in the labor market, driven by information technology and globalization, have created new requirements for knowledge and competences to fill job vacancies. This necessitates the development of innovative tools that provide timely information, facilitating better alignment between workers and employers while promoting the modernization of educational and vocational training systems. To address this challenge, an innovative software was developed that, through web scraping techniques, extracts data from Ecuador's leading job portals, such as Computrabajo (<a href="https://ec.computrabajo.com/">https://ec.computrabajo.com/</a>, 2023), Multitrabajos (<a href="https://www.multitrabajos.com/">https://www.multitrabajos.com/</a>, 2023), and Encuentraempleo (<a href="https://encuentraempleo.trabajo.gob.ec/">https://encuentraempleo.trabajo.gob.ec/</a>, 2023). The software employs an advanced methodology of text-mining techniques that can be applied in various contexts to generate relevant variables. It utilizes transformer-based architectures, such as all-mpnet-base-v2 and Sentence-T5-Large, to standardize the classification of economic activity sectors in job postings according to the Classification of Economic Activities (ISIC 4.0) and occupational categories based on the International Standard Classification of Occupations (ISCO-08). Once the occupation information was standardized, the competences and knowledge associated with each were incorporated, using the European Skills, Competences and Occupations classification (ESCO) as a reference. To complete the database, given that approximately 51% of the job postings lacked information (not present in the database) on the required experience and 41% on the level of education, the Extreme Gradient Boosting (XGBoost) ensemble algorithm was implemented to impute the missing data and recover all the missing information. The results revealed a high concentration of vacancies in positions related to technicians, mid-level professionals, salespeople, and managers, primarily targeting individuals with a high school diploma and, in most cases, at least one year of experience. Ultimately, this tool is established as an essential and cost-effective resource for the country, not only for its ability to improve the alignment between job supply and demand but also for its potential to complement and expand the current statistical sources on the labor market. The data obtained provide an opportunity to effectively guide the updating of vocational training systems, facilitating the creation of new training programs. Furthermore, these data can serve as key inputs for curriculum reform in higher education institutions, aligning their programs with the current labor market demands.</p>

Enhancing Labor Market Intelligence in Ecuador: A Framework for Generating, Standardizing and Analyzing Job Demand Data

期刊论文
发表日期：

2025年4月22日

查看全文

Diego Del Pozo-Villafuerte,
Andrés Villacís-Miranda

摘要

The rapid changes in the labor market, driven by information technology and globalization, have created new requirements for knowledge and competences to fill job vacancies. This necessitates the development of innovative tools that provide timely information, facilitating better alignment between workers and employers while promoting the modernization of educational and vocational training systems. To address this challenge, an innovative software was developed that, through web scraping techniques, extracts data from Ecuador's leading job portals, such as Computrabajo (https://ec.computrabajo.com/, 2023), Multitrabajos (https://www.multitrabajos.com/, 2023), and Encuentraempleo (https://encuentraempleo.trabajo.gob.ec/, 2023). The software employs an advanced methodology of text-mining techniques that can be applied in various contexts to generate relevant variables. It utilizes transformer-based architectures, such as all-mpnet-base-v2 and Sentence-T5-Large, to standardize the classification of economic activity sectors in job postings according to the Classification of Economic Activities (ISIC 4.0) and occupational categories based on the International Standard Classification of Occupations (ISCO-08). Once the occupation information was standardized, the competences and knowledge associated with each were incorporated, using the European Skills, Competences and Occupations classification (ESCO) as a reference. To complete the database, given that approximately 51% of the job postings lacked information (not present in the database) on the required experience and 41% on the level of education, the Extreme Gradient Boosting (XGBoost) ensemble algorithm was implemented to impute the missing data and recover all the missing information. The results revealed a high concentration of vacancies in positions related to technicians, mid-level professionals, salespeople, and managers, primarily targeting individuals with a high school diploma and, in most cases, at least one year of experience. Ultimately, this tool is established as an essential and cost-effective resource for the country, not only for its ability to improve the alignment between job supply and demand but also for its potential to complement and expand the current statistical sources on the labor market. The data obtained provide an opportunity to effectively guide the updating of vocational training systems, facilitating the creation of new training programs. Furthermore, these data can serve as key inputs for curriculum reform in higher education institutions, aligning their programs with the current labor market demands.