The cross-correlation of an acoustical signal transmitted into various rooms, allows the reflective and acoustical properties to be modeled. Using these models, different rooms can be classified and distinguished from each other. In this paper, we present a low-complexity and fast algorithm to classify different rooms and spaces. In contrast to other methods, our method requires minimal hardware: An Android device. Using machine learning algorithms, we model several types of spaces and use these models to classify new test spaces. Our method is validated and is shown to work with approximately 89% accuracy. © 2020 IEEE.