A couple of weeks ago, I hunkered down in my home office, refrigerator fully stocked, preparing for my work day. I checked into “Snowpocalypse 2015” along with some fellow millennial friends on my Swarm app as we prepared for Winter Storm Juno, which was anticipated to be one of the worst storms in New York City’s history.
I think that these weather phenomena have become quite fascinating from a social perspective. A far cry from the quiet, cozy snow days I experienced as a child, news and social sentiment have the potential to explode as each storm approaches. Information (whether fully accurate or not) is available through a single click. One of my early realizations of this was in 2011, when I experienced the effects of an unusual 5.8 Richter scale earthquake with an epicenter in Virginia. It was the first earthquake that I’d been able to feel, and was subtle enough that while the idea of an earthquake crossed my mind as a possibility, it seemed equally plausible to have come from the construction workers drilling outside. As the vibrations seemed to pass and I was deliberating whether or not to leave my apartment, I scanned Facebook and immediately saw my feed populate with questions from friends asking the same thing. Instant validation.
Seemingly quiet conversations on the U.S. East Coast on the Friday preceding Juno ballooned into a national dialogue over the weekend. Even though the South, Midwest and West Coast would remain unaffected, social activity about the storm, dubbed with hashtags and portmanteaus such as #Snowmageddon2015, reached a significant level of intensity in these regions too (check out this video on Fast Company to see the full effect).
Of course, as everyone who followed the storm knows at this point, the “historic snowstorm” didn’t quite turn out that way for New York City.
This was not to say that information was lacking – we have more data and models than ever before to predict the weather. The dilemma for decision makers was that while most models gradually lessened their predictions about New York City’s snowfall as the weather system approached, the European model, considered the best in the industry, predicted far more. Besides, citizen deaths could have resulted in a major PR crisis.
As consumers of these massive volumes of data, we want to believe that all of these insights are going to give us clear answers. Data can indeed offer powerful answers and help us make informed decisions to manage risk, but Winter Storm Juno also serves as a cautionary tale. Not every behavior or weather pattern can be predicted 100 percent, and there can be consequences ranging from trust issues to significant economic impacts when leaders have to make a call based on an outcome that is less than guaranteed. It is also a reminder for all of us to avoid the common trap of our own biases – it can be tempting to use whatever data looks most attractive for telling the story that you want to tell, and be done from there.
At IBM, we view data as the new natural resource and believe that next-generation analytics technology and expertise is essential to help us make the most of it. Many companies and governments are indeed tackling this head on, including in the realm of weather issues. The city of Rio de Janeiro, for instance, is beset by frequent violent storms, which cause destructive floods and landslides on its many heavily-populated hillsides. IBMers collaborated with the city to predict the weather down to the block level, and up to 40 hours in advance, through Deep Thunder, an application developed by IBM Research. The app uses high-performance computing, physics and a combination of topographical data, information gathered from global weather data from the U.S. National Oceanic and Atmospheric Agency and many other sources. In a similar vein, IBM’s cognitive system Watson becomes smarter the more questions it is asked, enabling practitioners across a wide array of industries to enhance their decision making.
What does the future hold? Some of the new paradigms we can expect in big data are more rigorous approaches to extracting and analyzing all data (not just what fits within narrowly defined expectations), new executive leaders focused on maximizing the company’s return on data, and ever-improving insights driving business decisions. It will require great resources and agility to get to this future state, but I believe that we are at a major tipping point. One trend that we are seeing within organizations is the emergence of a new role, the Chief Data Officer, as an agent of change with a broad mandate to guide this transformation.
As for the weather, this may not be the last winter storm of 2015, but we can definitely expect to see more advanced modeling and predictive analytics to aid leaders in tackling these issues around the globe in the years to come.
In the meanwhile, if you tend to be more superstitious, you can always turn to Punxatawny Phil who predicted six more weeks of winter on Groundhog Day.
Related:
- IBM100 Deep Thunder
- More big data and analytics posts on the IBMCAI blog
