Introduction
Hurricane Sandy was one of the greatest natural disasters ever to hit the United States. At the time, phone service was very limited due to cell phone towers being knocked out. But of those who could establish contact, how many people called the emergency hotline? What were the emergencies being reported? Where and when were these reports being made? To convey all this data properly, a map and several charts were necessary. They also had to be placed on a timeline to show users in real time what happened when.
Materials
The main tool used for this presentation was TableauPublic, an application designed to help visualize data. For the purposes of this exercise, datasets were obtained from a sample list provided on OpenRefine’s Wiki Page. The data was then cleaned using OpenRefine. It should be noted that not all of the data provided ended up being used in the final product. There were far too many categories to properly sort through, so it was important to be selective regarding which kinds of data would be relevant.
Methods for Visualization
The most pertinent information for the presentation are as follows:
- The volume of calls made to 911.
- When those calls were made.
- What those calls were reporting.
- When were these complaints resolved.
To accurately present the timeline, each of the graphs were spread out into separate pages. Each page represented a day, with the range being October 28th, 2012 to November 11, 2012. This range is the extent of the data presented in the original document, and it also serves as a gauge of judging when the hurricane hit and when life returned to the status quo.
The most prominent graph was a map of the New York area. Different zip codes on the map were highlighted to show the volume of 911 calls at any given time. Below that are a bar chart and a box and whisker chart. The bar chart represents the top ten complaint types per borough, and the box and whisker chart showcases how many of those cases were resolved.
The last chart used in the presentation was a simple line graph showcasing the number of calls made to 911 throughout the days specified in the timeline. This serves as a reference point for the other data presented in the first two graphs.
Results
The results garnered from this information actually defied expectations. Despite phone lines being destroyed and communication being limited as a result of the storm, 911 still received a high volume of calls starting on October 29th, 2012. Most of these calls came from more suburban areas, most notably Brooklyn and Queens. Less surprising, those areas received an overwhelming volume of damaged tree reports. In fact, there was actually an unexpected spike of damaged tree reports made on November 8th; well after Hurricane Sandy had hit. This could be due to phone lines being recovered, allowing for more people to contact the emergency hotline.
In more urban areas such as Manhattan and other parts of Brooklyn, the most prominent form of complaint was in regard to heating. These reports were scarce early on and towards the end of the timeline, but in the middle of the week, they managed to rival the number of damaged tree reports. Surprisingly, electrical reports did not make it into the top ten, and thus are not showcased on the graph.
Also, the user can utilize this data to pinpoint exactly when Hurricane Sandy hit, as well as when things started returning to normal. At the very beginning of the timeline, the most frequent use of 911 came in the form of noise complaints; specifically in regards to loud parties. The day after, however, those complaints dwindled while damaged tree reports skyrocketed. At the end of the timeline, on November 11th, the overall volume of calls was minimal, and the most frequent report regarded heating. This shows that while circumstances began to normalize at this time, lingering effects of the storm were still being felt. The box and whisker graph further reinforces this point, as it shows that certain complaints were able to be addressed a lot faster than others.
The map compliments this data by showing the specific zip codes from where those calls were made. Once the calls began to increase on October 29th, most of those calls came from zip codes which were close to the water. This trend continued as the days wore on, even after the storm had passed and most of the damage was already done.
The full report itself, along with all the pages, is displayed below:
