Positive matrix factorization (PMF) is widely used for apportion the source of particulate matter. However, the inherent model has some subjective aspects which should be reduce to ensure the robustness of the result. To do so, this study developed a systematic method, by performing tests on the input and the result validation. Finally, we proposed recommendations for input selection and result validation. A Python package is developed, providing advanced tools for input preparation, validation and visualization results.