TPS 641: Policies, interventions, communication, Exhibition Hall, Ground floor, August 28, 2019, 3:00 PM - 4:30 PM Metadata provide the meaning and enrich the data collected in a data warehouse; and they are a fundamental input to transform them into new knowledge that supports decision making. Specification and structuring of a data warehouse is a challenge for epidemiological analysis that considers that information is incomplete, that it includes variables with specific values and it omits relevant aspects such as its unit of measurement, range of allowed values, measuring instrument, source of the data, nature of the variable, measurement period, etc., which are considered a priori by interdisciplinary research teams. This work presents the experience of building a base metamodel that has been considered a complete and explicit framework for the implementation of a data warehouse for the discovery of knowledge about air quality and health through data science techniques allow him to make decisions on environmental health for the 10 municipalities located in the territory. Understanding that a metamodel is independent of the technology, we presents the physical implementation of the data warehouse in the PostgreSQL database management system whose license is free and supports the manipulation of large volumes of data; and that allows the integration of databases from different sources as environmental information (daily air pollutants concentrations, dispersion models), meteorological data and health data (daily counting of deaths and morbidity events, and synthetic indicators). In addition, it's presented how the statistical software R can access the information and some epidemiological and machine learning models are recommended for the mining of data of air quality and health useful for the evaluation of strategies and public environmental policies in the region. This tool arose from the interaction between environmental public authority of the region – Area Metropolitana del Valle de Aburrá- and Universidad de Antioquia, in view of the need for data integration and environmental health monitoring.