Seasonal forecasting puts weather data to work for your business
Using seasonal weather factors in your proactive strategy requires that you understand what things are knowable about the weather, both from an analytical perspective and a predictive perspective.
Beginning with an analytical understanding of the effects of past weather provides a basis for predictions about the future effects of weather. Moreover, understanding external demand drivers—via historical weather data, social media data and sales data—holds the key to understanding specific market-level weather influences on consumers. Indeed, many people don’t realize just how accurate weather forecasts and weather data are. For a good overview of technological advances in meteorological prediction, see Nate Silver’s article “The weatherman is not a moron”—and discover that weather forecasting provides some of the most objective and scientifically sound information available to business planners.
Market-level analytics can help us measure the influence of weather on consumer behavior—consider, for example, how fall weather influences buying behaviors. Such analysis can provide an accurate post facto perspective for benchmarking and explaining sales, both internally and externally, and can also serve as a basis for predicting future changes in both the short term and the long term. When forecasting the weather from a business perspective, consider the following four time frames:
- Recent past
What happened? How did weather shape the business? What weather variables, by market or region, affected the business most materially? Quantify such effects.
- Near-term future (0–10 days)
How might the trends identified in the analysis affect the business? What tangible or probabilistic changes to the business could drive additional sales or margin?
- Medium-range future (1–3 months)
How would the most likely macro-level climate trends affect the weather? How would such effects change consumer behavior, both from how it is during a typical year and from how it was last year?
- Seasonal planning (3–12 months)
What changes do we expect next year based on what we know about the analytical and historical effects of weather on the business, taking into account the difference between last year’s weather drivers and the most current historical averages for our location?
How does seasonal forecasting work in practice? Take a look at one forecast from 2014 that unambiguously predicted mild weather for the holiday season—and then contrast the prediction with what actually happened as recorded on CNBC and reported on The Weather Channel's Forecast Factor blog and measured by the National Weather Service.
To learn more about predicting seasonal weather, attend IBM Insight 2015, scheduled for 25–29 October in Las Vegas. Listen as IBM and The Weather Company demonstrate how to use weather data packages and IBM data science tools—such as Apache Spark as a Service in the Bluemix cloud—to address urgent weather-related business challenges by building powerful smart data applications for meteorological forecasting. Additionally, discover how to use Apache Spark as a Service to hack your analytics-related business challenges.