UFO Shape Dashboard in R Shiny
A week or two ago I started to play around with the NUFORC dataset. From the beginning I knew that I would want to create a few interactive dashboards with this dataset. So that's what I did.
My first dashboard is pictured below. You can use the dashboard to select a US State to see the distributions of UFO types that have appear in that state over the past 30 years.
UFO Shape Dashboard
NOTE The plot below takes a few seconds to load up, please be patient.
Just select the US State from the dropdown box on the left and see the counts of each type of UFO represented by bar lengths on the right.
R Shiny vs D3 for Dashboards
I used R's Shiny framework that basically provides an interactive data visualization experience based on R packages. You can essentially turn your work in R into a visualization. You can include R packages like ggplot2.
So you get all these powerful features and yet you have loads of customization options. In addition to using the R universe of packages you can treat your Shiny app like a web page with custom styling and full control over the layout. Not only that, but you can use R's data pipeline features to get real data into your visualization.
The other option I considered was using D3 which is a Javascript library that provides a language and framework for creating visualizations. D3 provides even more flexibility than Shiny, but you only really get basic features in D3. At it's core, D3 is a way to bind data to visual elements. So you can feed D3 in a sequence of numbers and each visual element (say a bar) will bind to a number. Bigger numbers have bigger bars. You also get basic scaling and other features needed for charts.
The D3 workflow is very much a building block experience: you start with simple elements and slowing add more and more elements. This allows you to build any kind of visualization that you want. However, it also means that basic plots like the one I have above take much more effort to produce.
So for now I will stick to Shiny, but I may revisit D3 in the future for a hobby project. Or if I want to create something that is more infographic-like I might considering pulling out this tool.
More UFO/UAP Visualizations Coming Up!
One of things I would like to do next is to find a way to search and then present a case interactively. Like I would want to see all the recent sightings in my hometown and I would also be able to cleanly see the details of the case. Where did the sighting take place, what is the full story.
Another thing that might be interesting is to do text analysis of some kind, although that might be better served as an article.
Finally, it could be interesting to see how reports appear geographically over time. You might be able to detect UFO "flaps" or otherwise see patterns. Do reports cluster in time? Do they spread like a virus?
What would you like to see? Comment below and I'll see what I can do.