Number Picture is a product of the work that went into a yet-to-be-published book which attempts to catalog every possible chart. There are almost 500 different chart types presented within its pages and many have not been seen before.
Motivation behind the project
In everyday life we are presented over and over again with a handful of chart types for disparate information. By now pie, bar and line charts have become part of our visual language but with such vast amounts of information to be communicated there is need to expand our visual vocabulary and express data more eloquently.
Some of the existing literature on the subject delves into the processes of creating visualizations. Other authors elaborate on the principles of good visualization or the semiology of visual cues and what they signify.
However, nobody starts at the beginning and enumerates exactly which charts we have at our disposal. This is the logical first step to take before attempting to provide comparative advice about which charts to use. We can achieve this first step by combining the available visual cues together in every possible way and this is what this project aims to do.
Visualizations essentially are collections of shapes and lines which vary from one another to show differences in their underlying data. These variations in the properties of the shapes or lines we call Visual Variables. Some common ones are:
- Position along a scale
In 1984 two scientists William Cleveland and Robert McGill provided a ranking system for the the visual variables listed above. Based on prior research they theorized and verified that underlying data in charts could be more accurately read with certain visual variables than with others.
This meant that they could tell beforehand that, all things being equal, a bar chart would be more accurately interpreted than a pie chart. This has large ramifications because it means that not only can we enumerate the possible charts at our disposal we can also rank them by effectiveness.
Currently we are limited in the ways we can visualize information by the software at our disposal. However, this is not the root of the problem - it is symptomatic of the fact that we don't have a coherent way of thinking about how charts are constructed. A classification system if you will. We only can think of charts as charts and not as collections of smaller components that, when combined, produce the actual charts.
The project is an attempt to formulate a classification system that will give us a common way of talking about and imagining charts of all types. A foundation that we can build upon in order to have more sofisticated software and a richer vocabulary for visualizing data.
A project by Finn Fitzsimons - data visualization enthusiast, software engineer. Book design by Boris Bonev.