In this project, we propose data analytics to identify the grade of glioma tumor. The process requires data collection, pre-processing, segmentation, application of machine learning methods, and interpretation of the results. The main methodology of this project is multimodal machine learning, and the main contribution is the fusion of different medical data modalities for diagnosing glioma. Our main objective is to identify the appropriate modalities (semantics) along with MRI and design an efficient deep neural network architecture for glioma grading. This model enhances the results’ accuracy, and we will extend this study to identify other types of tumors and injuries in the body.