In our upcoming book Everydata: The Misinformation Hidden in the Little Data You Consume Every Day, we have an entire chapter on some of the challenges in being a sound consumer of forecasts. One of the most prevalent sources of everydata forecasts is the weather report.
Meteorologists have a tough job—trying to forecast the dynamic systems in weather patterns. Many things go into the science of meteorology, and forecasts are based on a series of complex models of the weather. There are two common ways to think about forecasts:
- Often times people think about forecasts as purely deterministic—will it snow or will it not snow?
- The reality is, most forecasts are more useful when viewed as probabilistic—what are the chances of the DC area getting any snow? 1 inch? 2 inches? 3 feet?
The last few days the DC area has been gripped by Snowmageddon 2016 fury. Here's a tweet from our local weatherman Chuck Bell from Storm Team4—notice he is giving RANGES of snowfall amounts.
One of my favorite weather websites is the Washington Post's Capital Weather Gang, which does a good job of presenting forecasts in a probabilistic way. Here is an excerpt—notice phrases such as "Confidence remains high" and "still time for the storm tracks and impacts to change a bit."
Finally, here is a link to the latest forecast from Jason Samenow at Capital Weather Gang. Oh, and if you are in DC, based on my own personal forecast, I would be strategic about when to go to the supermarket over the next few days because there will be a run on it no matter what weather actually comes! If you can, stay home, read our blog, and sign up for our newsletter instead!