Building an Effective Data Dashboard: Results from Maine’s Maternal and Child Health Data Visualization Project

Wednesday, June 7, 2017: 10:30 AM
440, Boise Centre
Cindy Mervis , University of Southern Maine, Portland, ME
Erika Lichter , University of Southern Maine, Portland, ME
Fleur Hopper , University of Southern Maine, Portland, ME
Emily Morian-Lozano , CDC/CSTE Applied Epidemiology Fellowship Program, Atlanta, GA

BACKGROUND: Maternal and Child Health (MCH) data tables have been produced on a regular basis in Maine, but not in a way that was effective for many stakeholders to use. The aim of this project is to go beyond a series of data tables to more effectively communicate data to non-scientific audiences, including health department staff, other state and local government staff, clinicians, community partners, and other interested individuals.

METHODS: Maine was selected to participate in the Association of Maternal and Child Health Program’s (AMCHP) Data Translation Technical Assistance Project. A team from Maine worked with experts in data communication to design a web-based dashboard. Indicators were selected primarily from the Maternal and Child Health Block Grant. The dashboard, including all web pages, maps, figures, and tables, was created using SAS software.

RESULTS: Maine’s Title V MCH Program, in collaboration with the University of Southern Maine, AMCHP, and Evergreen Data, created a preliminary web-based dashboard, including figures, tables, and interactive maps, to present state and county level data on key MCH indicators. Figures were created using best practices in data visualization, such as having meaningful titles, using color to highlight key findings, and minimizing clutter. Using these techniques resulted in a visually appealing display of information that can be used at the local level for decision making. By using SAS software, we were able to standardize and automate the creation of tables, figures, and maps across counties and the state. This presentation will highlight effective data visualization techniques, some of the challenges of using these techniques, and decisions made as the dashboard was built.

CONCLUSIONS: Maine’s participation in the AMCHP Data Translation Technical Assistance Project facilitated the development of a web-based dashboard that utilizes best practices for data visualization. We found that effective data visualization can be done using SAS, which dramatically shortened the time needed to create web pages, including maps, figures, and data tables, and will allow Maine staff to easily and efficiently update the dashboard as additional years of data become available. Development of the dashboard required a high level of SAS programming skill and the automated approach to figure creation introduced some limitations when applying best practices for data visualization. Maine’s new MCH data dashboard will be a useful tool for displaying place-based MCH data and highlighting disparities at the state and county level.