This project is aimed at early prediction of Parkinson’s disease through computerized analysis of online handwriting samples. While a number of clinical evaluations are carried out to diagnose the Parkinson’s disease, most of these tests are effective only once the disease is at a relatively advanced stage. Studies have shown that analysis of handwriting can be used as a valuable tool for prediction of Parkinson’s disease at very early stages. In the proposed research, we have employed raw signals captured by a digitizer tablet while attempting several writing and drawing templates, to compute various kinematic, temporal, and pen-based attributes. In addition to these online dynamic attributes, static offline visual features are also extracted using state-of-the-art Convolutional Neural Networks (CNNs). The developed tools allow practitioners to collect handwriting and drawing samples of subjects and employ a combination of different types of features (online and offline) for the analysis and prediction.