Visualizing community exposure and evacuation potential to tsunami hazards using an interactive Tableau dashboard Completed

USGS research for the Risk and Vulnerability to Natural Hazards project at the Western Geographic Science Center has produced several geospatial datasets estimating the time required to evacuate on foot from two tsunami evacuation zones (standard and extreme) traveling at three travel speeds (impaired, slow, and fast walking speeds) for the Island of O’ahu, HI. Tabulation of O’ahu resident and employee counts by region, community, and the estimated travel speed necessary to reach safety within 15 minutes serves as the final dataset for conclusions. These data are useful for emergency managers and community planners to plan for tsunami evacuations, but are often difficult to serve using traditional static maps and graphs. Interactive dashboards are a popular way to display multi-dimensional datasets as they allow users to click and hover over graphics in the data. To provide easy access and promote deeper understanding of our evacuation results to O’ahu emergency managers this project explored visualizing these data in an interactive, online format using proprietary data visualization software called Tableau.



Principal Investigator : Jeff Peters

Cooperator/Partner : Jamie L Jones, Kevin D Henry

    The project team chose Tableau as the publishing platform because the capabilities advertised under a basic license matched what we needed to accomplish our goals, the license was relatively inexpensive, and as of the project start date, Tableau had not been thoroughly explored for USGS science purposes. Previous projects building an interactive dashboard from scratch proved time consuming and required a staff of talented programmers. By using a software package like Tableau, we hoped to save time and money while still visualizing the evacuation data in an interactive, online format to improve accessibility for the users. The project team consulted with the lead O’ahu emergency manager on dashboard design in order to provide the most effective user experience and provide a deliverable that was customized for their tsunami evacuation planning needs.

Accomplishments

The accomplishments for this project are described below.

• Evacuation results data were formatted for displaying in Tableau. Completing several tutorials was necessary to learn how to format data to be Tableau friendly, especially to link datasets and graphics which allows interactive features to be utilized.

• Two base maps of travel speeds (one for the standard evacuation zone and one for the extreme evacuation zone) were created using Mapbox for display in the Tableau dashboard map. Mapping capabilities of Tableau are limited to only one simple shapefile, so Mapbox was necessary to display the travel speed data spatially within Tableau. We were able to overlay community boundaries and link that to the evacuation data within Tableau which provided the capability we wanted, to auto-zoom the map when a community or region name is clicked. There is a potential for more sophisticated mapping within Tableau by incorporating javascript, but that was outside the scope and purpose of the project (to build the dashboard with little to no code).

• The elements listed above were combined with bar and line graphs of evacuation data, completing the goal of creating an interactive dashboard of Oahu tsunami evacuation results. Custom graphics capabilities in Tableau are limited so some design elements required learning many “workarounds”, as they are called in the various online help resources.

• The Tableau dashboard was embedded into a USGS webpage within the Risk and Vulnerability to Natural Hazards project website, including a help section on how to use the dashboard and links to background information, related publications, data-releases, and this project’s CDI Sciencebase and bitbucket pages. This accomplishes the goal of making the USGS Oahu evacuation dashboard, and accompanying information, easily accessible to Oahu emergency managers and the public. https://geography.wr.usgs.gov/science/vulnerability/Dashboard_3.html When interacting with the dashboard locally on your PC it is very reactive with little wait time for the graphics to update when clicked or hovered on. However, when interacting with the same dashboard through the website, it slows down significantly, which is disappointing.

• The time needed to become familiar with using Tableau through online tutorials and blog searches, as well as time used to design and develop our dashboard, was recorded. A comparison of the development of our Tableau dashboard and a hand-coded dashboard is included in the final presentation. This comparison was originally intended to be a cost-benefit analysis but we found that the vast variability in functionality of any interactive dashboard to be the major cost component, and due to limitations of Tableau dashboards, it is not appropriate to compare these development methods strictly based on cost. Ultimately, we think Tableau dashboards are useful for displaying USGS science, but it is necessary to consider the limitations and whether they could prevent completing project goals.
  
   

  Caption for attached image

  An image of the O’ahu Tsunami Evacuation dashboard. https://geography.wr.usgs.gov/science/vulnerability/Dashboard_3.html Users can interact with the graphics by clicking and hovering over community names and the graphic to find more information.