NOTE: This event was originally planned to be on Monday 14th Oct but has now moved to Tuesday 29th Oct.

At the heart of statistics is the desire to understand causal relationships, but the perennial problem of spurious correlation limits us. In some cases disentangling the causal from the correlated is literally a matter of life and death.

Problems like climate change and logistics require an understanding of spatial and temporal dynamics to solve. A "traditional" Data Science approach generally fails in these situations and can lead to a misunderstanding of the data and therefore potentially catastrophic decisions. By using visually intuitive spatio-temporal graphs we can simultaneously solve difficult problems and begin to disentagle the merely correlated from the causally linked.

In these talks we will show a simple and intuitive approach to constructing spatio-temporal graphs from data. We show why and how you would do this, with reference to some real world success cases.

So, join us on Tuesday 29th October at 6pm-7:30pm in Queens Building 1.69 for the first part of a four-talk series by Robert Nicholls