Click. The television screen turns on; the five o’clock news is just starting. What’s in today’s agenda: weather, politics and business? Over the hour the news hosts give information about what days you will need that umbrella, the outlook of an election and the stock markets projected changes. It is likely that all of these things, the weather, politics and stocks are just random information but there is an underlying relation between them all. All the information given was based on predictions.
What makes these predictions, an area in statics called forecasting? Forecasting is a statistical method of making predictions about the future based on previous data. Unfortunately, they are not able to predict everything and are not always right, as the predictions have to be based off of previous data. Regardless, forecasting has become an invaluable and integrated tool for people’s daily lives, as well as keeping society better aware and prepared for the future.
A sign of Forecasting. Image from http://www.picserver.org/images/highway/phrases/forecasting.jpg
For instance, let’s take a look at the devastating Hurricane Katrina. In 2005 New Orleans was hit hard by the deadly hurricane, affecting thousands of people. It may have been a lot worse however if it was not for forecasting! Forecasting was able to pick up on hurricane Katrina’s change of direction, moving from the Florida Panhandle to New Orleans. This gave the residents of New Orleans a heads up compared to what would’ve happened without a forecaster, a very deadly surprise.
Now, from what has been described it may seem like forecasters are amazing; having the ability to predict many things without limitations, but sadly this is not the case and what researchers at the University of British Columbia have been looking into. Vincenzo Coia along with Natalie Nolde and Harry Joe, are all part of UBC’s statistics departments and have been working on developing a new method for forecasting. The researcher’s goal with this method was to produce a forecaster that is able to be more accurate and predict extreme values. The video below shows an insight into exactly what Vincenzo Coia’s methods were, along with some student views on forecasters.
A video about forecasters. Video from https://youtu.be/3ofwa8xTxNI.
Vincezo’s study focused on predicting extreme values, using Copula. Though this may sound intimidating it’s just a fancy word that describes a method of finding the relationship between two variables. Why focus on predicting extreme values? I know I would rather want to know what’s the worst thing that could occur, not just that it will rain for instance. Listen how Vincenzo used the Copula to overcome some challenges in the forecaster design and more.
Podcast of the challenges and difficulties faced in developing a forecaster. Audio from https://soundcloud.com/taleen-chen/podcast.
Forecasters are society’s version of a crystal ball, allowing us to ‘look’ into the future. Already they can predict a variety of events, from the weather to stocks, but are limited by their accuracy and to the presence of past data. With researchers like Vincenzo and his team, one day perhaps forecasters won’t be limited to past data, but break through their limitations, revealing everything a crystal ball could.
–Xi Chen, Hong Yi Li, Sabha Mozafarian, Tristan Jeffery
