Microdata versus Big Data: Computational and Inferential Challenges in Data Science

Intensive-data-driven research is empowering theoretical breakthroughs and high-tech innovations, enabling new methodologies in academic discovery, and offering new sustainable means to solve significant societal and economic challenges and understand the world. Data science is a very fast growing research domain which embedded the combination of computational (i.e. computer-intensive) and inferential (i.e. statistics-oriented) thinking, with the notion of Microdata to Big Data. Since the core theories in computer science and statistical science were developed separately, there is an oil and water problem to be surmounted in data science. For an example, the basic statistical theory does not have a place for runtime and other computational resources while core computer science theory does not have a place for statistical risk and inferential resources. As well, the most appealing challenge in Microdata is the simulation of detailed population characteristics, while in Big Data is the potential of personalised attributes. In this seminar, I will address some major computational and inferential challenges in data science with special focus to combining data, computation, and inferences – that is not only computing statistics or running machine learning algorithms, but also there must be a focus on reliability measures including standard errors and confidence intervals on any outputs. Hence the Bayesian thinking as a combination of cognitive science, computer-based computational flexibility and statistics could be the sacred methodology in data science for advanced level modelling. This talk will also illustrate some real world case studies from data science.