I discovered a dangerous phenomenon the other day over conversation with my co-workers Bret Victor and Glen Chiacchieri.
Bret was speaking of the requests he gets to release the source code for his prototypes, and his rationale behind not doing so — that it could lead to innovation halting, because the release creates a standard for what products should be.
I realized,
When an inventor creates a tool and makes it available to the world, the public has a tendency to simply accept the tool as is, use it, and stop thinking of new ways to do things. (thus halting development of the product).
Example #1: the Pie Chart
Pie charts are pervasive…
… And a poor way to represent information. (See Edward Tufte, The Worst Chart in the World, and Oracle’s Reasons Not to Use a Pie Chart)
A simple example illustrates that pie charts make it difficult to make comparisons between two quantities. See:
What if we represented the same information like this, instead? This illustration enables us to make direct comparisons between quantities.
Because pie charts cannot spatially fit all information, they also are pleasantly accompanied by a key (right-hand side), which attempts to illustrate what all the components of the pie chart designate.
Could we design another way to represent this information — one that doesn’t require you darting your eyes back and forth to understand the information?
What about this?
Okay. Maybe not as aesthetically pleasing. Perhaps a reason why people use pie charts is because all the circles look pretty. But I would argue that we should not sell ourselves short — can we not have a representation that is both intuitively functional, and aesthetically pleasing?
I re-designed a representation of the same information; illustrated below.
So what?
These are just pie charts. Okay, it’ll take a millisecond longer to process the percentages. Why does it matter?
I think we vastly underestimate how good the quality of an experience (in this case, understanding and exploring the world) could be, because we already have models in our heads of what the medium is currently like.
One step further: Understanding Bayesian Probability
Another way to view this idea is through the lens of Bayes’ theorem. The traditional example is the cancer testing scenario, where you’re presented with a series of probabilities.
You are typically asked:
Assuming you have a positive test, what is the chance you have cancer?
When calculated out, the number seems much lower than expected.
Some of us simply accept that our brains do not intuitively grasp probabilities, but I’d argue that statistics is only unintuitive because we don’t have proper representations.
How would you design another representation of Bayes’ Theorem?
More to come.