Data collection in raw format will require an effort for standardization, preprocessing and monitoring (including alerts). This work foresees the development of a set of (Python) scripts to query a database and process the corresponding data, rendering it ready for further analysis. It is also expected that the student is able to complete a set of preliminary analyses on the processed data, namely in the form of statistics and visual representations on each dimension under study, and the application of *unsupervised learning* techniques to unravel specific patterns or *supervised learning* methods based on multimodal data to predict the patients' functional scores.
Ideally, the end result should be fed into a visual dashboard, e.g., based on Streamlit, that can be used by the physicians collaborating with the project to easily assess each patient's status, on any of these four dimensions.
Author: Matilde Silva
Type: MSc thesis