It’s Hard to Work Together (Especially with Data Scientists)
It may be surprising to hear, but creating a data visualization is not as simple as it sounds! When designing a data visualization, you will need to:
- Perform data exploration to learn more about your data and the features it utilizes.
- Design your visualization and plan out how you want to display your data.
- Develop and create your visualization using a variety of tools and software.
Each of these steps requires a lot of time and resources, which is why these projects are often done by a group of people, a team, in which each collaborator brings a unique set of Data Visualization skills. This is known as Data Visualization Design Handoff, where specific tasks are “handed off” to each member of the team with the intent of minimizing time and resources spent.
Sounds foolproof right? Unfortunately, that‘s not quite the case. In the research article titled “Data Changes Everything: Challenges and Opportunities in Data Visualization Design Handoff”, six data-specific visualization challenges are highlighted for handoff and design. Luckily, addressing these challenges allows for opportunities to create better handoff tools. I will summarize these challenges below.
For those who are interested in reading the research article, an embedded link to the article can be found at the end of this blog!
1. Adapting to Data Changes
More often than not, when working with a team, you will not have your complete set of data before the “Visualization Deployment” stage of your project as the data providers might send updated data later on in the process. This can lead to significant changes in the display of your visualization and even changes in the implementation stage. For example, changing a column name or a unit of measurement could break how you are parsing through your data.
It is important to note, however, that changes in data should not be avoided as they can provide corrections or additional data. The challenge is being able to cope and adapt to these data changes.
In the image above, a data update was performed, resulting in a compression of the original data and an overwhelming dominance of the newer data as the countries added primarily used coal. This completely altered the display and readability of the visualization as the original data was crowded into a much smaller slice, making it challenging to compare other energy sources.
2. Anticipating Edge Cases in Data
When creating an interactive visualization, it is practically impossible to test every possible edge case, or for interactive graphs, “the combination of interactive inputs that a visualization might receive.” Introduce handoff into the mix, where design and deployment are separated into different tasks, and the situation gets even more complicated.
The designers of the visualization may not necessarily have the knowledge or tools to test all the possible edge cases, so when the project is passed on to the deployment phase, users may encounter potential problems that were not noticed during the design phase.
At this point, making more changes to the design will add significant amount of work in both the design and development phases, which ultimately limits the possible ways these problems can be fixed.
3. Understanding Technical Challenges
As with edge cases, a similar problem arises when understanding technical challenges. Since Visualization designers focus primarily on the visual mappings of the graph, they may not be aware of the technical constraints that might occur during the development phase. This can cause problems for both the designers and the developers.
The developers may be tasked to implement a design that is simply not feasible, leading to more time spent in redesigning the visual. On the other hand, the designers may become wary or uncertain about proposing more unconventional designs, causing them to offer designs that may not be exactly how they envisioned.
4. Articulating Data-dependent Interactions
When creating data-dependent interactions, designers need to demonstrate several prototypes involving unique scenarios of the interactions in use. For example, a filtering operation reduces the amount of data being visualized, which can restructure/reshape the entire graphic. For this reason, a designer will need to illustrate these unique scenarios for the development team.
Like many of the previous challenges, this problem can require a lot of additional time and resources.
5. Communicating Data Mappings
One of the most critical challenges is effectively communicating data mappings throughout the Design Handoff. The mapping between the data and the visual itself is the “fundamental aspect of a visualization.” For this reason, the designer needs to be incredibly clear about the connection between the data and their design. However, this is easier said than done, as there are an infinite number of ways to misinterpret the intent of the design.
We can see an example of miscommunication in the image above. Here, the designer expected the data collected to be mapped to the size of the circles in the visual. However, “size” was not further explained, and the developers mapped the data to the diameters of the circles rather than the area of the circles.
6. Preserving the Integrity of Data Mappings Across Iterations
Surprisingly, it is quite difficult to compare implementations with design documents. On many occasions, “misinterpretations and misapplications of the data mapping go unnoticed before and after deployment.” The incorrectly-sized bubbles in the previous image is a perfect example of how difficult it can be to notice inconsistencies and mistakes in the implementation of the visualization. On top of all of this, numerous small changes are made for each iteration in the implementation process, making it increasingly difficult to keep track of what has been inspected and what hasn’t.
Hopefully now that you are aware of the challenges in Data Visualization Design Handoff, you will be ready to face them!
*All quotations and images are directly from the Data Changes Everything article.
