The emerging predictive health analytics provides great promise in reducing costs and improving health outcomes. However, most predictive models do not capture environmental exposures that impact health risk patterns in several chronic diseases such as asthma. This gap prompted the development of the exposome paradigm to improve health intervention by providing meaningful and understandable feedback on collected data. In this paper, we investigate a number of commonly used classification models applied in predicting health risks of asthma, given patients and environmental exposure datasets. We discuss the limitations of these existing models and propose a two-step approach of logistic and quantile regression, which provides a meaningful and comprehensive feedback for patients. The proposed approach uses a novel exposome assessment paradigm that utilizes the spatio-temporal properties of the data in the model training process and hence results in improving the accuracy of prediction. The quality of the proposed approach is extensively evaluated using real patients and environmental datasets.