FORESIGHT, Issue 24
- Stream Analytics For Forecasting by Patrick McSharry
Huge volumes of data now flow from on-line sources — the Internet, mobile telephony, weather sensors, and much more. With this data surge, there is an emerging need for computational algorithms to instantly process data streams, sift through large volumes of information, and extract and interpret knowledge. Patrick McSharry provides an overview of the challenges and opportunities arising from data streams.
- Forecasting Rounds of Golf by Scott Parrott, John Stamey, and Timothy Burcham
Back in the 1970s, Thomas Saaty introduced a methodology – the Analytic Hierarchy Process (AHP) – for obtaining and consolidating the judgments of panelists to aid in decisions of choice. AHP has been successfully applied in the forecasting of tourism, corporate earnings, and other areas. In this article, Scott, John, and Timothy describe the use of the AHP to develop judgmental adjustments to statistical forecasts of golf demand at a course in South Carolina. The statistical forecasts account for the trend and seasonal components of golf demand, while expert judgments were obtained on the effects of weather, economics, and local events, as well as course quality, availability, and pricing.
- Does a Presidential Candidate’s Campaign Affect the Election Outcome? by Richard Nadeau and Michael Lewis-Beck
Prevailing wisdom has it that campaigns don’t matter when it comes to forecasting U.S. presidential elections; the incorporation of direct campaign measures into statistical forecasting models does not appear to improve forecasting accuracy. Richard Nadeau and Michael Lewis-Beck now challenge that assertion, based on consideration of the quality and clarity of a candidate’s campaign. They find small but statistically significant campaign effects that can affect the outcome in an otherwise close election.
- A CEO’s Perspective on S&OP and Forecasting: An Interview with Phil Dolci
- The PollyVote’s Year-Ahead Forecast of the 2012 U.S. Presidential Election by Andreas Graefe, Randy Jones, Scott Armstrong, Alfred Cuzan
In 2004, Scott Armstrong, Alfred Cuzán, and Randy Jones launched the PollyVote to see if combining forecasts from different methods could improve the accuracy of election forecasting relative to individual forecasting methods. Scott had previously reported evidence that combining nearly always reduced forecast error below the typical individual method. As you’ll see in this article, Polly has performed up to and perhaps beyond expectations. Now she looks a year ahead (as of this writing) to predict the outcome of the U.S. presidential election. She thinks it will be close.You can read more about the origin and computation of the PollyVote in an article we printed in the very “first issue of Foresight: Cuzán and colleagues (2005).
- Outrageous Global Forecasts by Daniel Altman
In his latest book, Outrageous Fortunes: The Twelve Surprising Trends that Will Reshape the Global Economy, Daniel Altman looks at factors including geography, culture, and government policies to make predictions about the global economy in years to come. Here is a summary of what he has to say.