 This paper examined the performance of various machine learning algorithms in predicting the prices of apartments in Colombia. It compared the accuracy of linear regression, regression trees, random forests, and bagging. Additionally, it studied the effects of certain text attributes on the predictive power of these algorithms. Finally, it identified the most important attributes and explained how they relate to the overall goal of understanding the market. The results showed that random forests and bagging were the most effective algorithms, while text attributes did not have much influence on the predictive power of the algorithms. Furthermore, for attributes were found to be highly correlated with the price of the apartments, but their contributions to the predictive power were minimal due to the stability of the results across samples. This article was authored by Jorge Yvon Perez-Rave, Fadying Denzales Echivaria, and Juan Carlos Correa Morales.