Why do we need to learn more than one visualization tool to create beautiful data displays?
R is great for data analysis and static data visualization, but the default charts and graphs produced from R’s ggplot2 package needs refinement. For example, the basic output includes a lot of chartjunk, and lacks clear labeling. R’s new ggvis package produces graphs that inherently apply many more of the accepted data visualization design principles and leverage some interactive components of shiny. It’s easy to use Adobe Illustrator or a vector graphics editor to refine your charts. Python is great for data cleaning and data manipulation. Python’s matplotlib package for charts is a very powerful tool for plotting. The charts generated in Python also need refinement from a vector graphics editor. Read more about modifying charts using Adobe Illustrator. Users have a lot more control of the graphical elements of each chart in Python, but this requires more lines of code.
In summary, software is not magic. The most important thing is that your charts reveal meaningful insights and have a clear takeaway. Read more about the top five data visualization errors caused by software.
Dr. Kristen Sosulski develops innovative practices for higher education as the Director of Education for the NYU Stern W.R. Berkley Innovation Lab. She also teaches MBA students and executives data visualization, R programming, and operations management as an Associate Professor at NYU’s Stern School of Business.
Kristen’s passion for technology and learning sciences converges in all facets of her career, inside and outside of the classroom. Follower her on Twitter at @sosulski and learn more athttps://kristensosulski.com. Stay connected and join her #datavis newsletter.