You will be greeted on arrival as to which room we are in.
6:00 - Food and drinks
6:30 - Will Dracup - Using computers to find new blood tests
7:00 - Caterina Constantinescu - Data Scientist @ The Data Lab - Generalised additive models applied to tourism data
Generalised additive models applied to tourism data
In this talk I will discuss some of the functionality offered by R package `mgcv`. I will cover two generalised additive models (GAMs) built using this package, and applied to time series data measuring the monthly number of visitors at two historic sites in Edinburgh, Scotland: Craigmillar and Edinburgh Castles. These GAMs achieved a close description of the observed data, and largely respected the patterns of seasonality and other particularities that this time series data exhibited. One of these models was devised to forecast future visitor numbers, and used the data to predict itself. A separate explanatory model was also constructed, and included various additional predictors (considered to be plausible drivers of tourism activity), e.g., temperature, hotel occupancy data, economic factors etc. Relevant Google search trends were also included in the prediction, so I will briefly discuss the related package, `gtrendsR`. Finally, supporting functionality from packages `itsadug` and `mgcViz` will also be discussed.