(Read Time: About 3 Min.)
Could the total distance traveled by a NASA space shuttle have reached Mars? How far did each shuttle go in relation to our solar system? As an inaugural project to get our digital projects studio thinking and managing a project, we used a dataset that recorded the distance traveled by each space shuttle launch from 1981 – 2011 and worked to visualize the total miles traveled for each shuttle as a hypothetical race to the red planet.
For this quick project, we chose a csv dataset that was already clean and structured according to discrete categories including date of launch, total hours and miles per mission, and number of crew. We selected the categories that would be most useful for our visualization and added categories for the New York Times Headlines and their dates of publication.
You can see our data here.
We decided to use Processing for this particular project, an open source programming language that’s well suited for electronic art and design projects because of the way it teaches programming concepts in a visual way.
For this particular project, it made the most sense to use the Model-View-Controller programming pattern. This framework served as the best way to translate our raw spaceship data into a visual, moving display. In addition, MVC also aligned with our division of tasks and made team collaboration possible.
As a result, we were then compelled to ask questions like: How will we structure data into objects and classes so that it aligns with MVC? How will we show data through the screen, or user interface? How will we coordinate all of these pieces together? Ultimately, in Processing, we addressed these challenges using different draw cycles for each shuttle and a synching pattern corresponding to the mission timelines.
As a team, we each separated into specialized roles in order to tackle visualizing this data-set. Pairs worked together to figure out how to create a “movie” in Processing, how to visualize the shuttles and planets, and how to pull corresponding metadata (in the form of New York Times articles) associated with each launch.
Over time, some key design decisions had to be made, including what our visual layout would look like, how shuttles would be “mapped” onto the solar system, and where associated headlines would appear during our movie. We used whiteboarding and sketches to brainstorm ideas and move forward with these decisions.
After three weeks of collaboration, we “launched” (pun intended) a final version of our space shuttle movie! This end-product showcases team talent in visualizing the shuttles, programming their movements, and highlighting news stories corresponding to each.
One key insight was determining when work should be automated versus manual when it comes to analyzing big data. With data visualization, there are often “rote” tasks, such as data-cleaning, that are necessary to execute before beginning any programmatic work.
In a similar vein, while our dataset was not particularly large or “messy,” there was the related task of collected NYT headline data which called for a “mixed-method” approach (i.e. using both automated and manual techniques to accumulate this information). While the headlines are conveniently stored in the backfiles database of the New York Times, the amount of information available was overwhelming. One strategy was to manually sift through headlines following each launch date, while another was to create a Python script using the New York Times API to automate the process. Although the automation was theoretically faster, it required time for coding and still produced an overwhelming response that had to be manually selected. In the end, considering our project timeline and how data had to be filtered by “relevance,” selecting the headlines manually was the preferred method; a certain xkcd comic helped affirm this decision: