Before you move the needle — A Data Visualization Design Framework!
Why this article?
I conducted a recent talk in #design4india, a NASSCOM initiative to create a significant impact on the cause of design in India.
The talk was with reference to — Data visualization design framework which could be used by designers to understand what they are designing for, and how to better prioritize and to select the right combination of visualization and interaction techniques, provided by the framework.
I received many requests to share the framework and the article is a continuation of the talk.
here is the video
I have started the article with “Is Designing for Data Visualization tough?”, for readers who attended the talk on the framework and would like to use it, dive straight into the framework — scroll till you reach, ‘The Framework and how to use it’ section.
Are creative people bad at Math
In my foundational years designing for Data Visualization — which goes back to 2007, during the Rich Internet Applications era, the biggest hurdle i faced was with regards to, understanding statistics, or how data visualization is represented.
Maths seems to be difficult, statistical concepts like — Mean, Mode, Correlations — sends shivers down the spine. But if we looked at these concepts from what they are really mean and what problems they solve — they end up becoming tools in our arsenal.
Its never too late to learn maths, or get better at statistics. I have part attempted in uncovering the most common ones among these — including Targets, Baselines, Unit of measures etc.
Evolutionary, Corrective and Retrospective design
As the article, “Iterative Rework”, suggests Evolutionary design is good for design outcomes, but generally Corrective and Retrospective design work hits efficiency and outcomes, these take way the time and result in rework for consumers of design direction — the product team.
Evolutionary design is an important part of uncovering patterns of problem and solution areas which helps us move the needle further, which are not visible on looking at data alone but by some amount of immersive effort.
Any framework, method or philosophy would be encouraged — which could help us drive more efficiency for design.
Bubble charts vs. Bar graphs
Here is a real life instance, on why you should consider — the learning curve and plan better when designing.
While designing for a complex visualization for a procurement application — i realized a bubble chart, which can encode up to 4 dimensions, was the best fit. It was efficient, looked really cool and saved a lot of space.
I continued with validating the visualization with end-users. to my surprise the end-users did not get it, on the first or the second pass. The variants i had made — in form of bar graphs were working better.
My initial read was, I probably should not include bubble charts in my design. My long term read of this, consider the learning curve.
The data visualization framework — includes a method to understand and rationalize — when you can take chances with the learning curve and when you cannot — using the prioritization grid.
Badly Defined Problems and Scope
There are no Bad Problems, but badly defined problems.
The problem of articulation, on what are we trying to solve. Without a clear definition we are left in lurch trying to understand the scope and filling in the gaps ourselves.
While designing data visualization projects, especially for the enteprise, this potentially increase cycles, iterations which are of corrective and retrospective nature.
The data visualization framework — attempts to include some aspects that force answer On some high level questions, which can help prompt more questions — whether it be about the users, the environment, the measurements and others.
Why teams don’t build visualizations which seem to work fine with end users and are innovative
There is a need to involve development teams, product management, and other stakeholder, early in the cycle along with end users. This is an way touches the essense of design thinking.
Involving teams early gets a better understanding of perspectives, on the feasibility and visibility aspects, which should not limit but help expand our understanding.
What is the cost (Investment) and what are the likely returns
- TBD
The framework and how to use it
The framework is a result of years of designing for enterprise data visualization projects in Supply chain, Manufacturing, Finance, Healthcare, and Procurement. Its is a result of finding patterns that are repeated in some of my projects, while keeping the solutioning open.
The framework consists of two important apscets of information grouping and visualization . The scope and object mapping, and the prioritization grid.
Scope and Object Mapping
The scope and object mapping begins with the scope — this includes aspects like — The organization user level (Who), Frequency of Use (When), Outcomes (Why), workflow (How), Activity type(What), and Complexity (Derived).
Let’s use an example of a conference, to help walk us through the framework. Assume we are asked to build a data visualization for the conference, which includes the CEO, the conference manager, and the Operations executive, and they possibly have different objectives, and frequency of usage for those objectives.
Role Layers (Why)
The first question we ask, “Who are we are building the Data Visualization Experience for?”, The answer might be in terms of a Persona, or a reference to the hierarchy in the organization they would belogn ti. The important thing, might be to ask, does it belong CXO, Manager, or the Operations — classification.
The CEO — possibly has very little time, and can only spend little time, and mind space on the data visualization, a simple KPI number might suffice, as they might have to look at hundreds of other things as well, without going too deep.
The Manager on the other hand, might spend reasonable time on some tasks to uncover the problem area, and assign it to an executive to solve the problem.
While the Operations executive, needs to be both reactive and deep in terms of root cause analysis.
This is different from a Persona, which goes much deeper. A Persona, that you might have created from user research, might be helpful to answer this question.
The frequency of use (When)
The next, one could be — how often would they use the visualization? Once a year, every day, or every quarter.
The Manager operations possibly would use it on a daily basis, when the conference is on, and on quarterly basis when they plan the next conference.
The other question’s we try and answer include — Outcome, the Activity type, the Workflow etc.
The Objects and the UI Patterns
It’s now important to collect aspects which will help us create some visualizations — for instance identifying objects, related for the objects including — unit of measures, baselines, targets, projections etc.
What are objects
Objects are often described as Measures, and dimensions in different visualization topics, I wanted to avoid an overlap and hence, coined the term.
Objects can also be stated as entities that appear in visualization for instance, how many participants attended a conference, or what was the cost of the conference.
Here the Participant, is the object — the count is the unit of measure, we could have a target for the number of participants, we want to attend, or the baseline — comparison with other conferences, or for the matter thresholds — how many would be too less, or too many for the conference to run successfully.
What are UI Patterns
It might also make sense to learn about the UI design patterns that can be leveraged. Including the chart elements, the filters, etc. The idea is not to limit us, but to inform us what are the interaction design components, that easily available to us.
Scope and Object Mapping Cards
These cards are a result of collating single row items led by the workflow — or scenario title. We end up putting together, the Who, the when, the why, and the other aspects . These cards will help us get the necessary information to create the visualization or the experience.
Applying the scope and object mapping cards to create visualizations
Scope and object mapping cards are relatively useful methods to explore visualizations, within what we know. As we learn more, we can update and add more value.
Here are two examples, of how object mapping cards help. Both examples are not exhaustive, but a look at how you can create more, informed by goals, objects, etc.
In the first example, we discover How “Topics of interests across the years”, this has the potential for a conference planner, to understand, what should be the ideal and relevant mix of topics for the conference. The card stack displayed on the right, summates information which helps us move forward in this regard.
The visualization composition is a design/ creative process, but the card stack is making it more tangible, and also helping us understand — what is possible with the metrics we already have.
Using the same card stacks, we could create another visualization, Where we compare “Cost vs. Participation”, here again the Objects such as Cost and Participation with their, unit of measures are helpful.
Prioritisation Grid
The prioritization grid is the aspect, where we plot the acquired knowledge on a grid — comparing the problem area on generic to detailed, and frequent to once a year, aspect.
Assuming we continued doing the exercize, and plotted more scenarios for each user, we might end up with a prioritisation, like displayed below.
The CEO, looks in frequently, and need information on a very generic nature. the Manager operations on the other hand, looks at information which might be once in a year, or a quarter, but needs real granular data for planning. While the operations executive — would look at it on a granular basis and frequently — while the conference is ensuing.
A combination of different card stacks and methods would get us a better refined model. The prioritisation grid — would help drive conversations better, and also be more aware of the scope we are dealing with.
Whats Next!
I am hoping to evolve this framework further. I encourage you to leave me comments, use the framework, and put it together in a way — in helping us designers move the needle further for Business and social outcomes we care about.
I am thankful to the design team in VMware, who helped me add value to the framework and bringing this forward. If you would like to learn more about VMware Design, please visit, vmware.design.
