IEEE International Systems Conference · Apr 6, 2026
Exploratory Analysis of Dementia Risk Factors Using Distributed Data Systems
By Cail T, Saha PR
Dementia is one of the leading diseases that causes disability among the elderly. Usually, dementia is caused by a combination of demographic, educational, lifestyle, and medical factors. An analysis of these risk factors is crucial for identifying priority areas for intervention to reduce dementia cases globally. Some analysis has been performed in the past to identify common risk factors for dementia based on data sources with a uniform structure. However, the evolution of different computing platforms enforces the accumulation of data in multiple modes. Moreover, the sources of the datasets may not be in proximity. Accommodating such modern paradigms of data in dementia risk factor analysis is paramount. In this research, we present a distributed data system which is capable of integrating multiple publicly available data sources and demonstrating the relationship between lifestyle and dementia. The data sources can be in different proximity and have semi-structured data with different schemas, which includes individuals demographic and lifestyle information. Our proposed distributed data system efficiently handles the semi-structured data files (e.g., CSV or JSON) and produces a correlation analysis. The analysis focused on descriptive statistics rather than predictive modeling, highlighting modest associations between dementia and several factors such as education, smoking, and alcohol consumption. We implemented the system in the Hadoop distributed database framework (HDFS) with an integration of PySpark and MongoDB, and experimented with two large public data sets from Kaggle. The results suggest that higher educational attainment may offer a protective effect, while lifestyle shows moderate associations. Our proposed system is scalable, and preliminary experiments demonstrate its ability to support a reproducible exploratory and descriptive analysis of risk factors for dementia.