Visualization of health indicators: utilizing data mining techniques and statistical analysis for effective comparison of user profiles

Abstract In this paper, we describe an integrated framework which incorporates data mining and statistical methods leading to the development of a visualization tool to facilitate users in understanding and exploring health indicator data. Due to the increasing cost of providing health care and decreasing participation in physical activities, motivating individuals to self-monitor themselves and according to improve health is an important case to consider. Here, it is essential to emphasize that the developed system will not replace professionals in health care who are expected to have hands on serious cases. Our target is to provide a framework flexible enough to allow for periodical or daily follow-up of personal indicators and hence may raise some concerns leading to early warning to go for serious check in the clinic. By using visualizations, we hope to aid a variety of potential end users in understanding and exploring different aspects of their health. In the back-end, the system employs some data mining and statistical methods for effective prediction to be used in articulating the current state of the interacting user. User interactivity is important because we are trying to allow users to use the system for a personal approach to improve health. Using data mining techniques, we will present the user with a variety of different views to compare their personal profiles with. By making active comparisons, we hope to provide a context for users to understand how his/her health levels compare to others within the same age range or location. We concentrate on Canada as a case study, though the proposed framework could be smoothly adapted to fit any other case for which the relevant data are available. The conducted user study demonstrated the applicability and effectiveness of the developed framework.