As part of my PhD research on Large Scale Data Mining for Healthcare, I am developing a system called Computational Healthcare.
Computational Healthcare indexes and aggregates data from millions of healthcare visits. Using novel aggregation strategies it simplifies data analysis, replication and dissemination process. Researchers can conduct millions of different studies in few clicks. While Physicians, Medical students and Patients can use Computational Healthcare to understand outcomes and future complications.
Computational Healthcare is developed in collaboration with Radiology and Anesthesiology departments at Weill Cornell Medical College, Cornell University.
I am a co-founder of Temporal Health. Temporal Health aims to make information derived from large healthcare databases accessible to a broad audience.
Using EraseImage users can perform image segmentation and background removal using superpixel algorithms. It works completely in client-side JS. It is implemented using Angular, FabricJS and Superpixels.
An experimental version uses neural networks to perform segmentation and erasing. Training is performed inside browser and trained models can be saved and reused for similar images. This version also works entirely in browser and is implemented using the ConvNetJS library.
I am interested in developing tools for mining large publicly available datasets (E.g. the Common Crawl dataset ~100 Terabytes ), using cheap cloud computing resources such as AWS spot instances.
Additionally I have been analyzing data derived from sitemaps of Pinterest for advertising and user acquisition.