Detection of Diabetic Neuropathy - Can Visual Analytics Methods Really Help in Practice?

Abstract

Visual analytics (VA) methods are valuable means for supporting the detection of diabetic neuropathy, the most common longterm complication of diabetes mellitus. We suggest two strategies for strengthening reliability, reproducibility, and applicability of dedicated VA methods in practice. First, we introduce a novel workflow visualization that shows activities together with metadata and produced output, facilitating a guided step-wise analysis. Second, we present a tailored user interface that integrates various VA tools, unifying access to their functionality and enabling free exploration for further assisting the medical diagnosis. By applying both strategies, we effectively enhance the practical utility of our VA approach for detecting diabetic neuropathy.