Big Data analysis
Much of our research utilises sensory devices to measure individuals’ exposure to air pollution to explore links between environmental factors, daily activities and health outcomes. The very large datasets produced in such studies requires specialist data analysis techniques.
We apply the four major elements of Big Data analysis, referred to as the ‘Four V’s’: volume, velocity, veracity, variety, with a particular emphasis on veracity and variety. We focus on the development and utilisation of state-of-the-art time series and pattern recognition techniques to extract spatiotemporal patterns in individuals’ daily activities (e.g. indoor activities: cooking, sleeping; and outdoor activities: walking, transportation on a bus or in a car) and discover their associations to individuals’ health records and/or symptoms diaries.
Our objectives are to:
- identify patterns in large data streams obtained from environmental stationary and mobile sensory devices;
- investigate links between spatiotemporal sensory patterns and health records and heterogeneous data sources;
- identify the quality of information in data streams