Data Visualization Explained (Part 2): An Introduction to Visual Variables

article in my information visualization sequence. See the earlier article: “Data Visualization Explained: What It Is and Why It Matters.”

So, now you’ve realized the foundational concept of what underlies information visualization and why it’s an integral part of the information science ecosystem. (If you’re not acquainted with this, make sure to try the article linked above.)

As we mentioned within the earlier article, the core concept of information visualization is discovering an efficient strategy to characterize information of varied sorts in a visible method.

The important thing underlying idea which makes this illustration work is named a visible encoding channel. A visible encoding channel is successfully the means via which numerical, textual, or another type of information is translated into a visible mark. One of the best ways to think about it’s as a visible function equivalent to all or a part of your information. Efficient information visualizations typically use a number of visible encoding channels for various elements of the information.

On this second article, we’ll dive into the small print of visible encoding channels and acquire some follow breaking down a fancy visualization into its part components. This can put together you for designing your personal visualizations within the close to future.

Introduction to Visible Variables

In his 1967 work, The Semiology of Graphics, French cartographer Jacques Bertin outlined seven “retinal” variables, named as such as a result of the human eye’s retina is delicate to them [1]:

1. Place (such because the coordinates on a graph)
2. Measurement
3. Form
4. Coloration hue
5. Coloration worth (lightness to darkness)
6. Orientation
7. Texture

Though Bertin printed his work a long time in the past, his visible variables stay a superb guideline for contemporary information visualization design. Within the early phases of growing a visualization, it’s good follow to evaluate the visible variables accessible and decide which of them to make use of for particular variables within the information.

This could be a complicated idea and is extra simply understood with an instance. The graphic beneath, typically thought-about a masterful utility of visualization, was designed and drawn by Charles Minard. It depicts Napoleon’s failed invasion of Russia.

It is a simplified and translated model of the map to ease readability; for the unique, see here [2].

What completely different visible variables are getting used within the graphic above? (Trace: There are fairly a number of.) As an train, get out a pen and paper and attempt to decide this your self. We’ll stroll via it intimately in a bit.

Maximizing Effectiveness of Visible Variables

The perfect visible variable to make use of for a selected visualization is dependent upon the information. Right here, we are going to have a look at three several types of information:

1. Quantitative: Numerical information with a pure ordering that’s appropriate for mathematical operations (i.e., it is sensible so as to add/subtract/multiply/divide particular person information values). For instance, wage and age are quantitative variables.
2. Ordinal: Categorical information (i.e., non-numerical information which may tackle a set variety of values) that also has a pure ordering. When you have ever taken a survey with reply selections akin to “Strongly Agree,” “Agree,” “Impartial,” “Disagree,” and “Strongly Disagree,” then you might have seen ordinal information in motion. Whereas mathematical operations on this information don’t make sense, varied values can nonetheless be ordered from “greatest” to “worst,” so to talk.
   • This additionally consists of variables which will have an order with out technically being “ranked,” akin to visitors gentle patterns.
3. Nominal: Categorical information which has no pure ordering. An amazing instance of that is colour. Whereas it’s doable to tell apart between completely different colours, they don’t have any pure sequence. (This additionally explains why colour is a wonderful visible encoding for nominal variables usually, as we’ll see beneath!)

Essential: Simply because a variable is a quantity doesn’t mechanically make it quantitative. For instance, zip codes are numbers, however they don’t have any pure ordering, nor can one carry out mathematical operations on them. Thus, zip code is a nominal variable.

The next desk, a variation of 1 designed by visualization specialists Jock D. Mackinlay and Stuart Card, outlines the effectiveness of various visible variables relying on the kind of information [2]:

Just a few key factors about these rankings:

• Place is the most suitable choice for all variable sorts. For instance, a bar graph with names on the x-axis and blood strain on the y-axis makes use of place for each a nominal variable and quantitative variable, respectively.
• After place, desirability modifications for every variable kind. That is necessary to know as a result of in case you are graphing a number of variables, you’ll finally have to make use of one thing apart from place as a result of it’s already getting used (often on a 2-D graph with two axes).
  • Size is an extension of place, however particularly helpful for quantitative comparisons.
  • Density and saturation are nice for ordinal variables, as your viewers don’t want to find out actual values—they simply must see the rankings.
  • Hue and form work effectively for nominal variables, making it simple to see categorical variations.
• Some choices are totally crossed out as a result of they merely wouldn’t make sense. For instance, form shouldn’t be a doable encoding alternative for quantitative or ordinal variables, as a result of there could be no strategy to examine portions or perceive orders.

Now, let’s stroll via an instance of learn how to break down visible encoding channels intimately.

Minard’s Map: Breaking Down the Variables

Let’s have a look at Minard’s map of Napoleon’s invasion collectively. Right here it’s once more for comfort. This instance is taken from Edward Tufte’s well-known visualization ebook, The Visible Show of Quantitative Info [3].

A cautious research of this map exhibits Charles Minard’s mastery of visible encoding channels as nothing in need of sensible. His visualization shows six completely different variables:

1. Geographic Location (Quantitative): Place is used to show the placement of Napoleon’s military on a 2-D floor (so that is technically two variables). The invasion started on the left facet of the map, on the Polish-Russian border. We are able to additionally see how at occasions, components of the military department off to completely different places as a part of Napoleon’s technique.
2. Geographic Location (Quantitative): See above.
3. Time (Quantitative): Trying carefully, we will see that varied cut-off dates are listed on the chart’s x-axis on the backside of the visualization. Once more, the place is used to show this variable.
4. Temperature (Quantitative): Temperature is plotted in relation to time on the chart beneath the map. Place is used but once more, this time on the y-axis.
5. Variety of Troops Remaining in Military (Quantitative): The width of the form shifting throughout the map represents the variety of troops in Napoleon’s military. It’s clear that because the invasion progressed, Napoleon’s military turned smaller and smaller. They finally returned to Poland with solely 10,000 residing troopers out of an preliminary 422,000.
6. Course of the Military’s Motion (Nominal): Coloration is used to depict the route wherein the military strikes at varied positions. The beige/tan colour (white within the simplified picture we’ve above) signifies the military’s motion towards Moscow, and the black colour signifies its retreat again into Poland.

In his ebook [3], Tufte refers to Minard’s map as presumably “one of the best statistical graphic ever drawn.” Finding out it might encourage us to plan intelligent methods to encode our personal information visually.

Remaining Ideas and Trying Ahead

With this second article, you’ve realized the foundational concept behind visualization design: visible encoding channels. As you mirror on what you’ve realized, hold the next key factors in thoughts:

• The selection of visible encoding channel can typically make or break a visualization. You might need a fantastically designed graphic, but when the visible encoding channels are laborious to interpret, your viewers gained’t know what you’re attempting to say.
• Place reigns supreme for all variable sorts, however there may be restricted house in a 2-D setting. As such, consider carefully about which variables you show with place; they’ll typically be an important ones.
• Check out completely different designs! There isn’t a “one” excellent answer. Somewhat, it’s essential to revise and reiterate till you attain a passable level.

Within the subsequent article, we’ll discuss necessary suggestions for visualization design and the way methods have developed and expanded during the last a number of a long time. Till then.

References

[1] Semiology of Graphics, Jacques Bertin (translated by J. Ronald Eastman)
[2] https://ageofrevolution.org/200-object/flow-map-of-napoleons-invasion-of-russia/
[2] Readings in information visualization: using vision to think (Card, Mackinlay, and Shneiderman)
[3] The Visual Display of Quantitative Information, Edward Tufte