Making Analysis Interactive, Shareable, and Good-looking: How I Use R Shiny in Civis Platform
By Hayley Arader
The most challenging part of being a data scientist isn’t always getting results from data, but determining how to share those results in a clear and actionable way. At Civis, we work with stakeholders on all ends of the technical spectrum — from quantitative analysts to high-level decision makers — and making our analysis accessible and clear can be tough, especially at the code-reviewed and version-controlled high standard that I’m used to.
Interactive applications solve this problem by providing a framework for data scientists to communicate lots of information effectively and efficiently. Decision makers, who may have limited data experience or time, can use these applications to interact with high volumes of data in an approachable and actionable way.
That’s why I’m excited to share that R Shiny is now supported in Civis Platform. Shiny is an R package which allows R users to directly turn their analysis and plotting scripts into an interactive application. It’s a powerful tool that enables data scientists to turn their code into a visually informative and interactive platform, allowing teams of data scientists and decision makers to work in concert. The ability to launch Shiny apps in Civis Platform is great for me and my fellow data scientists because it fits right into our workflow and makes sharing insights in a compelling way so much easier.
I recently worked with a national organization that had data in lots of different formats from partners all around the country but didn’t have a way to use the various pieces in tandem. For each geographic area, we needed to show background information, area-specific predictions, and, more importantly, we needed to visualize how those dynamics applied to the entire country. With Shiny, I was able to quickly consolidate the many moving parts and give decision makers the power to compare and filter analyses so they could make informed decisions about where they should allocate time and resources.
Shiny has been transformative for both the Civis team and our clients in its ability to provide an effective communication channel between data scientists and decision makers. Decision makers can interact with a higher volume of data in a digestible and actionable interface, ultimately empowered to make data-driven decisions for their organizations.
Shiny is also a game-changer for data scientists because it empowers us to build the best possible product and to share our insights more quickly and easily than ever. We can choose the best way to present insights to prevent information overload for decision makers: data scientists love our algorithms and models and optimizations, but we also love it when our work is understood and used. Shiny lets us demonstrate the value of our work to folks across the technical spectrum and become an essential part of making decisions.
Best of all, Shiny fits seamlessly into a data scientist’s workflow. Many data scientists are already comfortable in R (and those who aren’t usually find it easy to pick it up). Writing a deliverable in code drastically improves the quality of both the workflow and end product by transforming the process into just another part of the data science workflow. Shiny makes it easier for data scientists to collaborate, streamline code review, manage version control, and to share code with our colleagues so they can adjust it and repurpose it for their needs.
Writing the deliverable in R is what allows us data scientists to reach our full potential and develop the best possible product without needing to spend any time learning HTML and CSS to build a full front-end application. Shiny allows us to present our data with tools and workflows we’re already familiar with, giving us more control and flexibility over our entire process, from the actual data science to the final presentation and consumption.
