Impact of Data in Visualization Design Handoff
Complex data visualization design projects involve collaboration between people with different visualization-related skills such as designers, developers and other collaborators such as data experts and project coordinators. Each of them contributes in specific phases of the project such as design, implementation and deployment. These members need to communicate clearly amongst themselves for implementing the project. For communicating clearly, it creates the challenge of handoff. Handoff refers to exchange of information between the project members working on different phases; it is the output of each phase of the project.
In this article, we discuss different stages of a complex visualization project, six data-specific visualization challenges for design specification and handoff and identify opportunities for future tools to overcome these challenges.
First, let’s look at the different stages of a visualization project:
· Project Conceptualization: This stage occurs on the client side before the design team is involved. The client provides the design team with dataset, vision and goals of the project. Handoff in this phase can be anything from a simple email to an involved workshop or a meeting between the client’s data experts and design and development teams.
· Data Characterization: In this phase, the members of the design team explore the dataset provided by the client and prioritize its characteristic based on the project’s vision. The handoff of this phase is summaries of data characteristics.
· Visualization Design: In this stage, a visualization concept is developed and approved by the client and then the final design is refined, polished and documented. The handoff of this phase is visualization design documentation.
· Visualization Development: This phase is led by the development team. They implement the designs handed off to them. The designer team’s role in this is answering questions of the developers, suggesting redesigns when issues arise, and confirming the implementation works as planned.
· Deployment and use: In this stage, the visualization is deployed for public use. The development team handles the maintenance and data updates. The design team only gets involved only when if the data updates contain outliers which are not supported in the existing design.
Now that we discussed the different phases of the visualization project, let us look at the challenges when designing with data in the project.
Adapting to Data Changes
Data updates can have cascading effects on each phase. Even when data change does not affect overall mapping, it might affect the implementation stage, particularly where server-side mechanisms for aggregating, loading and preparing data. For example, changing a column name may break existing data parsers.
Anticipating edge cases
As the name suggests, edge cases mean rare cases or outliers in data. It is almost impossible for the designer to expect all possible combinations of inputs that a visualization may receive. So, it is not easy to anticipate the values that can break the design or data mapping.
Understanding Technical constraints
Designers might not be aware of all of the technical constraints that occur during the development phase. They focus more on providing an appealing interaction experience which may have some technical constraints for the developers to implement. For example, these constraints can be designs like chart and graph combinations which are difficult to achieve for the developers.
Articulating Data-Dependent interactions
When ideating design interactions, designer needs to generate a variety of views showing visualization in many states. This extra cost can make interactions challenging to develop and communicate to the developers. For example, filtering operations reduce the set of data being considered. This may result in a change of position or appearance of remaining data.
Communicating Data Mapping
Implementing data mapping accurately needs detail and precision. But present design tools do not support precise and complete specification of data mappings. This difficulty in conveying intent can be because of data complexity, interaction complexity such as multi-dimensional and interactive visualizations.
Preserving Data Mapping Integrity across iterations
As there are multiple iterations in visualization projects it is difficult to compare the implementations against the design documents which will cause misinterpretation and misapplication. So, maintaining the data integrity from design stage to development stage is difficult.
Now that we discussed the challenges in this project. Let us now look at few opportunities to mitigate these challenges by using tools such as:
· Data characterization tools: These tools can help the designers better understand how the data has changed from one form to another and how these changes may affect the design of the visualization.
· Data ideating tools: These tools would ease ideating data driven designs. These tools will help provide accurate and easy to understand handoffs from designers to developers.
· Data mapping frameworks: These frameworks will support explicit communication between data structures and their graphical representation.
To summarize, data is hardly defined during the implementation of the project. Data related challenges occur all the way from design stage to deployment stage that include adapting to data changes, anticipating edge cases, articulating data-dependent interactions, communicating data mappings and preserving data mapping integrity across iterations in the implementation stage. These challenges point to several opportunities to create tools that directly support the visualization design and make the design process more robust, efficient and accessible to people in various specialized roles.
These challenges and opportunities mentioned above are anchored from own experiences of the members of visualization team who worked on five large multidisciplinary data visualization design projects conducted between 2012 and 2019.The main contribution of them is identifying the gaps between data characterization tools, visualization design tools, and development platforms that pose challenges for designer-developer teams working to create new data visualizations and identifying opportunities for future tools for prototyping, testing, and communicating data-driven designs, which may contribute to successful and collaborative data visualization design.
This blog is based on IEEE Transactions on Visualization and Computer Graphics 2021 paper Data Changes Everything: Challenges and Opportunities in Data Visualization Design Handoff by Jagoda Walny, Christian Frisson, Mieka West, Doris Kosminsky, Søren Knudsen, Sheelagh Carpendale, Wesley Willett.
