My team at SAP has been collaborating with the Laboratory for Computational Physiology at MIT which is part of the Harvard-MIT Health Sciences program. Physionet is the premiere publisher of anonymized research databases for medical research. We are using SAP HANA to help build the next releases of the their MIMIC research database. This is a unique use of HANA in that the clinical and signal information contained in the MIMIC dataset have a complex structure which requires very complex queries and signal processing when mining the data for clinical/treatment innovations in the Critical Care space.
We had a chance to share some of our innovations with a full house of data scientists at the annual Conference on Knowledge Discovery and Data Mining (KDD), a conference that provides premier forum for advancement and adoption of the “science” of knowledge discovery and data mining. We jointly taught a hands-on tutorial entitled, “Management and Analytic of Biomedical Big Data with Cloud-based In-Memory Database and Dynamic Querying“, in which we demonstrated some of the techniques we implemented to work with the MIMIC II database on SAP HANA. The data scientists were all given access to the HANA version of MIMIC II running on HANA One (HANA on the Amazon Cloud). Participants could follow the tutorial and try out the example codes on their own computer(s) simultaneous to the instructors.
None of the participants had ever been introduced to the performance benefits of in-memory computing or HANA, but they finished the tutorial quite impressed with the technology. We heard overwhelmingly positive feedbacks about the performance of very complex queries, the integration of analytical libraries, such as R and AFL, as well as the built-in visualization capabilities of HANA studio.
I have attached a few of the presentations so you can get a flavor of the type of things we are doing with HANA.
If you have any questions for feel free to email me at SAP.