Data science in sickness and in health

The relationship between statistics and data science is intimate yet awkward. 

On one side is statistical science: focused on parametric modelling and concerns about inference, leading to some of the abuses we see in much applied research.  On the other side is data science: putting computers to work on large datasets to mindlessly identify patterns or to obtain unprincipled predictions. These rigid separations are unhealthy; both camps make huge contributions to our understanding of the world, but often hold each other at arm’s length. However, their overlapping aims mean that each side can learn much from the other. 

As a group of statisticians, NASH appreciates the power of data science and this symposium is designed to help educate us and to produce interactions to facilitate a happier and healthier relationship between data science and statistics. 

Speaker presentations

Professor Peter Diggle
What is health data science? A biostatistician’s answer
 
Dr Dionisio Acosta Mena
Data science in healthcare and biomedicine: principles, practice, potential and pitfalls
 
Professor Stijn Vansteelandt
Machine learning for the evaluation of treatment effects
 
Professor Chris Holmes
Bayesian fitting and evaluation of complex models arising in modern health applications

Dr Jessica Barrett
Cardiovascular risk prediction using big data: a statistician’s perspective

This event is one of a series of Annual Events organised by UCL's Network of Applied Statiscians in Health.