A themed Forecasting workshop was the latest event organised by ExIStA.
This event was one of a number of themed workshops planned by the Exeter Initiative for Statistics and its Applications (ExIStA) and was held at the University of Exeter on 28th June. Forecasting was chosen as an ideal topic for an ExIStA workshop due to the broad range of applications that cover this area of statistics, and the variety of topics discussed during this event certainly did not disappoint!
Joana van Nieuwkoop-McCall presented her wave-height model and the statistical techniques she is using to verify its results (available online here). The model is being developed to provide forecasts of good weather windows on the Cornish coast to allow the deployment and maintenance of offshore wave energy infrastructure.
The second talk introduced the predictive risk model for emergency hospital admissions in Devon. Todd Chenore, from NHS Devon, explained that these models can reduce unplanned admissions by forecasting which patients are at high risk. Clinicians can then proactively work with these patients to reduce their risk of admission.
The Met Office provided two speakers; Michael Sharpe focussed on the verification of weather forecasts, while Andrew Colman presented his work on longer-term seasonal forecasts. Michael has developed a system to convert the hand-written shipping forecasts into a numerical forecast allowing forecasters to verify and improve them. This is the first time the shipping forecast has been verified since the BBC started to broadcast it in 1925 and UK sailors will be glad to know it has found to be very accurate! Andrew’s talk had an international perspective, as he described how he is working with the national met services in Africa to improve their seasonal forecasts using both statistics and weather forecasting models. These types of forecasts are invaluable for understanding how agriculture and water resources may be affected by the weather in the coming season.
Professor Jim Davidson presented his model developed to forecast UK retail food prices. This work was funded by the Department of Food and Rural Affairs (DEFRA) after concern over the large increases in food prices during 2007/08. James described the statistical techniques he has used to develop this model and what variables affect the cost of food.
The diversity of talks at this workshop highlighted the importance of forecasting in many sectors and the range of statistical techniques that can be used to build and verify forecasting models. What was particularly impressive was the range of benefits that the speakers highlighted including improved health and safety, NHS management and contingency planning.
Future themed events planned include a health workshop and a career-young event – look out for details on the ExIStA website.