Skip to main content

Using Online Applications with R to Share Surveillance Data

Public Health Problem (100 words)

Since 2009, the Cook County Department of Public Health (CCDPH) has created and disseminated weekly surveillance reports to share seasonal influenza data with the community and our healthcare partners. Surveillance data is formatted into tables and graphs using Microsoft Excel, pasted into a Word document, and shared via email listserv and our website in PDF format.

While disseminating data in this format has some advantages, it also has the following limitations: graphics are small and hard to read, minimal historical data can be displayed, graphics are static and cannot be manipulated, and there is little room to include spatial data.

Success Story Narrative (400 words)

To address the limitations in our static surveillance reports, in 2015, CCDPH set a goal of developing an online, interactive application that could serve as a companion to the report. In our vision, the application would display the same data available in static reports but also allow users to manipulate graphics (i.e. choose what previous influenza seasons to display), provide more information (i.e. exact values) when users hover over data points, and have additional space for larger graphics and maps.

The primary obstacle to executing our vision was a lack of expertise in web development and associated programming languages, such as HTML, CSS, or Javascript. In addition, there was no funding available to recruit outside resources, such as a web development company. However, in early 2017, staff from CCDPH began learning the R programming language and participating in the ISDS R Users Group. While exploring the R Users Group webinar library, CCDPH learned about the R shiny package – a tool for creating online, interactive web applications. We began exploring whether this tool could be used to develop our flu surveillance application. Despite having only six months of R coding experience when beginning, we were able to construct a functioning application that satisfied all requirements of our original vision between July and October of 2017. During this process, support from the R Users Group, in the form of webinars and specific technical assistance from the group chairs, was instrumental to our success, as well as the availability of other free, online tutorials. The final product used entirely open-source tools and therefore cost nothing to build or operate.

The online application debuted on October 13th, 2017 – in conjunction with the release of our first 2017-2018 seasonal influenza surveillance report. The link to the application was emailed to our flu listserv and posted on our website. It can be viewed here: https://ccdphcd.shinyapps.io/weekly_influenza_surveillance_beta/.

Outcomes And Impact (400 words)

As of February 23rd, 2018, 19 weeks after its debut, the application has been visited over 1,000 times. Since local influenza activity was elevated to “High” in late December, it has received an average of over 10 visitors per day. We have received positive feedback on the application from hospital infection preventionists, members of the media, and the general public. A feature unique to the application - an interactive map showing local influenza activity by zip code throughout the season - was demonstrated at the county’s GIS (Geographic Information Systems) Day and won a prize for “Best Application”.

Lessons Learned (400 words)

The number of visitors to the online influenza surveillance application has exceeded expectations and suggests that public health constituents are ready and eager to consume data in non-traditional formats. Making data available in online, interactive visualizations provides greater reach for public health data and messaging and demonstrates that health departments are technologically current. Our experience shows that health departments need not be constrained by funding or technical expertise to make applications of this kind. By taking advantage of open-source tools as well as the knowledge and expertise within the surveillance community, such as the R Users Group, CCDPH was able to achieve a goal that had previously shown minimal progress for two years.

Our application development is still in progress. We have taken suggestions from the user community throughout the influenza season and incorporated changes to the application’s content and design on an ongoing basis. Based on our experience, applications do not need to be perfect to be well-received by the community, so long as you continue to make improvements. The R code that powers our application is available on Github here: https://github.com/kb230557/Flu_Shiny_App. We would encourage others to use any portions that might be useful for their own projects and provide feedback on improvements that could be made to ours!

Submitting Author Name
Kelley Bemis
Submitting Author Title
Epidemiologist IV
Submitting Author Organization
Cook County Department of Public Health
Submitting Author Phone
708-836-8666
Submitting Author Email
kbemis@cookcountyhhs.org
Referenced File
Submitted by Anonymous on