Let's Talk about the Weather (and Analytics)

As Florida awaits Hurricane Irma's landfall, we look at how improvements in today's weather predictions rely on real-time big data and technologies like machine learning.

As islands in the Atlantic assess catastrophic damage and Florida residents get ready for the arrival of Hurricane Irma, many people are looking at data and satellite images about what looks like a record breaking storm. One popular data news site is saying that it looks like a worst case scenario.

Evacuations are underway in Florida, residents are boarding up their homes, and everyone is preparing their property as best they can against the damaging storm to come.

Weather affects everything. From the retailer who stocks snow shovels ahead of the blizzard and the pop-up street vendors in cities who sell umbrellas, to the insurance companies who now send alerts to their customers about expected weather conditions that could impact claims, businesses already know about the impact of storms and the importance of data and predictions about weather. Data feeds of weather conditions are now a part of the mix for many businesses and multiple firms offer such data and insights.

And in terms of predicting the weather, you may still hear people complain that forecasts are wrong. But there's more data available to help with predictions than we have ever had before. Now we have so many more sensors in so many more locations. We have satellite images. We are collecting data about constantly changing conditions from the ocean to the earth to the atmosphere. Weather is like the Cinderella use case for big data and real time data and machine learning.

And while there's so much more data and technology being applied to weather predictions now, there is also much more information available to the general public now, too. It's exciting for analytics enthusiasts and weather nerds like me who really enjoy talking about the weather and consider it to be much more than small talk.

For instance, you've probably seen the hurricane spaghetti and cone data visualizations that show the probable paths of Hurricane Irma. Or maybe you want to look at what the National Hurricane Center has to offer on the National Oceanic and Atmospheric Administration (NOAA) website. For instance, click through here for a list of the forecast models used to predict hurricane tracks and intensities.

Meanwhile, other sites may offer data and visualizations that provide a snapshot of the response to a major storm. For example, the web site offers a real-time view of flights, and as I'm writing this on the morning of Friday, September 8, there are still plenty of flights in progress over Florida. But there's a big empty spot to the north of eastern Cuba where a NOAA satellite image shows Hurricane Irma to be right now. (Of course, both these images will show different conditions when you click through the links. It's the nature of real-time data visualizations.)

You may also be wondering about how Irma stacks up compared to other hurricanes. Earlier this week, Scientific American put together a short article providing comparisons which you can find here. Harvey, which pummeled Texas in the last few weeks, is number 2 in terms of rainfall. Other metrics in the post include peak winds, lowest atmospheric pressure, and greatest diameter. Of course it's too early at this point to know how Hurricane Irma, which is still active, (or Jose behind it) will stack up against these record holders. But it gives us a benchmark of what to look for as Irma progresses.

Finally, not really data related, but here's one of my favorite images and headlines that relates to the recent Hurricanes: Harvey and Irma, Married 75 Years, Marvel at the Storms Bearing Their Names.

Editor's Choice
Brandon Taylor, Digital Editorial Program Manager
Jessica Davis, Senior Editor
Cynthia Harvey, Freelance Journalist, InformationWeek
Terry White, Associate Chief Analyst, Omdia
John Abel, Technical Director, Google Cloud
Richard Pallardy, Freelance Writer
Cynthia Harvey, Freelance Journalist, InformationWeek
Pam Baker, Contributing Writer