Given the high cost of collecting labour demand for skills information through surveys, the composition and dynamics of Colombian labour demand are relatively unknown. However, information regarding unmet labour demand can be collected from job portals with the implementation of relatively novel data mining techniques. This paper provides a descriptive analysis to start evaluating the data from job portals. Among the main results of my analysis of the vacancy database show that 1) job vacancies are concentrated in Bogotá, Antioquia and Bolivar; 2) most of the job positions require a person with at least a high school certificate; 3) most occupations in Colombia correspond to middle- (“Sales demonstrators”) and low-skilled occupations (“Kitchen helpers”); 4) the skills most demanded include “Customer service” (knowledge), “Communication” (knowledge) and “Work in teams” (competence). Thus, the vacancy database provided detailed, real-time and valuable information about the Colombian labour demand that, previously, it was not possible to obtain from other sources (e.g. household surveys). Moreover, these initial results suggest that the vacancy database is consistent, or at least it does not contradict itself or external data, such as regional GDP, population, etc.