Information Visualization: Principle and Methods

In a digital panorama dominated by huge knowledge and complex algorithms, one would assume that the common individual is misplaced in an ocean of numbers and knowledge. 

Isn’t it?

But, the bridge between uncooked knowledge and understandable insights lies within the artwork of Information Visualization. 

It’s the compass that directs us, the map that guides us, and the interpreter that decodes the mass quantity of information that we encounter each day. 

However what’s the magic behind visualization?

Why does one visualization enlighten whereas one other confuses?

At present, we’re going again to the fundamentals and attempting to grasp the basics of Information Visualization. 

Let’s uncover all of it collectively!

Mastering how you can inform a narrative effectively is likely one of the hardest abilities to grasp as an information scientist. If we test the time period Information Visualization in a dictionary, we discover the next definition:

“The act of representing data as an image, diagram or chart, or an image that represents data on this manner”

This mainly implies that Information Visualization goals to craft a narrative from the dataset, presenting insights in a type that’s digestible, interesting, and impactful. 

Information visualization, or making knowledge look good in charts and graphs, won’t appear as cool as stuff like machine studying. 

However, it is actually a key a part of what a Information Scientist does. 

In at present’s data-driven world, Information Visualization is just like the glasses that assist us see clearly. And, for these not well-versed within the language of numbers and algorithms, it affords an environment friendly means to grasp complicated knowledge narratives.

Any chart is at all times composed of two essential parts: 

1. Information Varieties 

I wager you’re considering of information as numbers, however numerical values are solely two out of a number of varieties of knowledge we might encounter. At any time when we visualize knowledge, we at all times want to contemplate what varieties of knowledge we’re coping with. 

Along with steady and discrete numerical values, knowledge can come within the type of discrete classes, within the type of dates or instances, and as textual content.

When knowledge is numerical we additionally name it quantitative and when it’s categorical we name it qualitative. 

So any displayed knowledge can at all times be described in one of many following classes.

As soon as we now have a transparent what sort of knowledge we now have, we have to perceive how you can encode this knowledge into closing charts. 

2. Encoding Data: The Visible Lexicon

Visible encoding is on the core of information visualization. It interprets summary numbers into graphical representations, a language we’re all fluent in.

Regardless that there are various several types of knowledge visualizations, and at first look, a scatterplot, a pie chart, and a heatmap don’t appear to have a lot in frequent, all these visualizations might be described with a standard language that captures how knowledge values are become blobs of ink on paper or coloured pixels on a display.

However… as you already should pay attention to…

There are literally thousands of methods to encode numbers!

There are two essential teams:

1. Retinal Encodings: From form, dimension, colours, and depth, these are components our eyes catch immediately. They’re inherent to the aspect.

1. Spatial Encodings: They exploit our mind’s cortex’s spatial consciousness to encode data. This sort of encoding might be achieved by place in a scale, an outlined order or utilizing relative sizes. 

With all of the beforehand defined encodings, we might use all of them in a single chart however it will be onerous for the reader to understand all the data shortly. Overloading a chart with a number of encodings might be complicated so 1 or 2 retinal encodings per chart is perfect.

All the time do not forget that much less is commonly extra, so at all times attempt to create minimalist and easy-to-understand charts. 

Consider it as seasoning a dish-a sprinkle of salt and pepper may improve it, however pouring all the salt shaker may damage the style.

So now… which encoding ought to one select? 

That, my buddies, is dependent upon the story you need to weave. 

So you can higher ask…

Whereas the visible arsenal at our disposal is huge, not all weapons are match for each battle.

Take into consideration what encodings are greatest for what sort of variable. 

• Steady knowledge variables, like weight and top, discover their greatest illustration in place on a standard scale. 

• Discrete ones, comparable to gender or nationality, shine when depicted by colours or spatial areas.

There are some causes behind the intuitiveness of some charts. And there are two essential theories behind them. 

1. Gestalt Principle 

Individuals who work with know-how generally overlook concerning the human aspect of issues. The Gestalt Rules are guidelines from psychology that designate how our mind sees patterns. 

A few of these guidelines assist us perceive why we group issues that look alike or discover issues that stand out. 

1. Similarity: Gestalt similarity means our mind teams issues that look alike. This may be due to their place, form, colour, or dimension. That is extensively utilized in warmth maps or scatter plots. 

1. Closure: Objects inside a border-like a line or a shared color-look like they belong collectively. This makes them stand out from different issues we see. We regularly use borders or colours in tables and graphs to group knowledge.

1. Continuity: When particular person components are related, our eyes assume they belong collectively. Even when they give the impression of being completely different, the road makes us see them as a gaggle. That is extensively utilized in line charts.

1. Proximity: We expect issues are in the identical group if they’re shut to one another. To point out issues belong collectively, put them shut. Utilizing somewhat area can assist separate completely different teams. That is generally utilized in scattering plots or node-link diagrams. 

So the Gestalt rules and their interactions are vital to contemplate when making visualizations.

2. The Precept of Proportional Ink 

In many various visualization eventualities, we characterize knowledge values by the extent of a graphical aspect. 

It’s common follow to make use of the phrase ink to seek advice from any a part of a visualization that deviates from the background colour. This consists of traces, bars, factors, shared areas, and textual content.

For instance, in a bar plot, we draw bars that start at 0 and finish on the knowledge worth they characterize. On this case, the info worth is just not solely encoded within the endpoint of the bar but additionally within the top or size of the bar. 

If we drew a bar that began at a special worth than 0, then the size of the bar and the bar endpoint would convey contradicting data.

In all these instances, we have to be sure that there is no such thing as a inconsistency. This idea has been termed the precept of proportional ink by Bergstrom and West.

“When a shaded area is used to characterize a numerical worth, the realm of that shaded area must be instantly proportional to the corresponding worth.”

Violations of this precept are fairly frequent when attempting to control knowledge, specifically within the well-liked press and on the earth of finance.

Related points will occur each time we use graphical components comparable to rectangles, shaded areas of arbitrary form, or every other components which have an outlined visible extent that may be both constant or inconsistent with the info worth proven.

A putting stability between aesthetics and performance is pivotal. Adhering strictly to rules like Bergstrom’s proportional ink, however not at the price of readability. 

And whereas some encodings could seem much less efficient, they are often chosen intentionally to make an announcement or evoke an emotion.

In our age of an ever-increasing movement of information, the significance of crafting significant visible narratives can’t be overstated. Particularly when attempting to speak our insights to non-data professionals. 

Good knowledge visualization isn’t nearly presenting numbers, however as a substitute attempting to articulate our knowledge round a narrative. Bringing our knowledge to life whereas telling tales, and forging connections between uncooked data and real-world implications and insights. 

As technologists and knowledge lovers, it’s our artwork, our language, and our bridge to the entire world.  Josep Ferrer is an analytics engineer from Barcelona. He graduated in physics engineering and is at the moment working within the Information Science subject utilized to human mobility. He’s a part-time content material creator centered on knowledge science and know-how. You may contact him on LinkedIn, Twitter or Medium. 

The post Information Visualization: Principle and Methods appeared first on AIPressRoom.