The late summer and fall of 2017 had more than its fair share of disastrous weather, particularly when it came to hurricanes making landfall.  In the midst of these storms and disasters, I was taking a course on Information Visualization where I learned to critique and analyze visualizations of many kinds.  Disaster-related visualizations abounded, including an interesting one in a post from the New York Times blog The Upshot discussing the costs associated with Hurricane Harvey and predicting those costs based on past disasters.  The accompanying chart demonstrated “billion-dollar natural disasters since 1980” and had a source to clear data, something not always found in disaster visualizations.  For the purposes of my class, I critiqued this visualization using the framework of “From Data to Wisdom” from Hunter Whitney’s book Data Insights: New Ways to Visualize and Make Sense of Data, which encourages analysis of the underlying data behind visualizations.

The interactive visualization uses data from NOAA (National Oceanic and Atmospheric Administration) National Centers for Environmental Information (NCEI).  This data is easily accessible and clearly sourced on the original post. Each disaster is clearly labeled with its cost and some disasters have accompanying PDFs with further explanations.  This kind of clear explanation is not as common in visualizations as you might think!

The chart shows a visual representation of each billion-dollar disaster in the form of a colored semicircle with the color indicating which kind of disaster (flooding, tropical cyclones, etc).  The semicircles are “sized proportionally to their cost in 2017 dollars.”  Years are along the y-axis, months of the year are spread along the x-axis, and the semicircles are placed in their appropriate date on the chart.  Harvey’s predicted-cost semicircles blink at the top of the chart in August 2017, highlighting its devastating potential.  This small segment of the chart comes from separate data not included in the NOAA NCEI dataset as the full costs are not yet known. The article discusses those financial predictions.  

Snapshot of an Interactive Visualization from The New York Times blog “The Upshot”

The data from NOAA and the Harvey estimates from other sources were interpreted for the article and visualization by New York Times graphics editor and reporter Kevin Quealy. He utilized the information publicly accessible from NOAA about billion-dollar disasters as a concrete comparison to the estimates being floated by different data modelers.  Quealy also makes conclusions in the accompanying article about the economic impact of these expensive disasters and speculates about the causes.  Overall, the NOAA data is presented in a straightforward manner with relevant suppositions and caveats.  

One complaint I have about the visualization is in the nature of the semicircles that made it difficult to hover over the desired disaster.  For example, you cannot hover over any disaster that appears behind Hurricane Katrina, as it envelops the years above it. This is visually appealing, as you can see how big in scale the cost of Hurricane Katrina was.  However, it is not practical in terms of accessing all the data presented.  You can go to the data source at NOAA and find what is lurking behind Katrina’s bubble, based on the month and year.  

If the question asked is, “How much will Harvey cost?” then this chart is not the answer. Past performance is not always an indication of future performance and the data is just not yet known.  Yet this chart succeeds in answering the following relevant questions: “How many billion-dollar disasters have there been since 1980?” “What was the financial magnitude of those disasters?” “Given these past disasters, how might Harvey’s cost fit in comparison?”  The conclusions from the visualization are concrete by showing the financial toll these disasters take.  Other comparisons and analyses could be made, but the overall visualization and data used are solid.

The next time you’re looking at a visualization in a news article, blog post, or presentation, take a deeper dive and check the source data.  Consider the questions the visualization is attempting to answer.  You may be surprised by what you find (or what you don’t!).

Sarah Davis is a Bilingual Youth Librarian at a public library in Oklahoma and an MLIS student at the University of Oklahoma-Tulsa.