What's Wrong With Chart Selectors?
When I started my visualization career in Deloitte Analytics, in beautiful Singapore, I was thrown in the deep end. Data Visualization was not my background and I was in a role that demanded I get up to speed real quick.
Drinking from the proverbial firehose, Dr. Abela's chart chooser was thrust in front of me as a best practice for picking the right charts. No matter how I approached it, I could not digest it in terms I could relate to.
“The chart chooser didn't click for me and it still doesn't”
Don't get me wrong, it works for many people. It also fills a void for people new to vizualization and gives them a place to start. So, this is not a critique of the venerable chart selector, more of my own limitations to make it work for me and feedback I receive with clients, on the speaking circuit and in my own training programs.
Chart selectors don't land for many people. Indeed, they can come across as complex and unapproachable.
My view on why they don't land is because the categories of charts are a step of abstraction away from how many of us think. Let's look at a human-centric approach do building a dashboard:
Understand who will be using the data. Persona creation. What are their goals and challenges?
What questions does the persona want to answer?
What data is available to answer those questions and how to iterate through the development cycles to answer them?
What visualizations does the persona need to see to answer those questions?
I'm a big believer in bringing the end user of a dashboard along for the journey. That means maintaining simplicity and having buy-in at key stages along the path. That includes chart selection.
Questions are a simple concept. Mapping those questions to charts is where is can get sticky. Sure, part of the role of a data viz professional is to help in chart selection, but let's assume we want a process where anyone can do this type of work. Indeed, the more people that can do effective chart selection, the more people will be adopting dashboards.
Going from questions and running them through a typical chart selector adds a degree of complexity that, in my view, is unnecessary and potentially limits storytelling and goes too deep into the weeds.
That's where chart patterns come in. Before I get into the details, let's look at a pros and cons.
Popularized by Andrew Abela, a chart selector breaks charts into several categories.
Chart patterns take in the wider business context and focus on chart interactions and relationships to get to a sufficient level of detail in order to take action on. It aligns with more human-centric thinking and answering questions rather than considering chart categories.
Chart patterns are more accessible and do not require technical understanding of chart best practices. The point of entry is through business context. For example, imagine this scenario:
Chart Pattern - Sample Scenario
Mid-management level employee with historical dependance on spreadsheets.
Supply chain operations for a computer game publisher. Lots of products.
How to I spot any issues in the supply chain across all products and caregories?
By taking typical business scenarios, we can determine what the ideal chart and chart interactions are, embed chart best practices and create a path to action.
In this example, we want an operational level perspective for spotting outliers in a supply chain process. There are several other scenarios that this pattern would apply to, of course, but this pattern provides a solid foundation to work from, at least as a starting point.
In summary, the reasonable approach would seem to be use both methodologies. Honestly, I never use a chart selector. I always start with a pattern and customize from there. Sure, experience helps a great deal, but I find the patterns to be a far greater accelerator for the novice and for working with non-technical business stakeholders.