1.

1 Introduction to data visualization

Sunday, 02 October 2022 8:06

• Types of data
• Benefit of visual representation of data
Summary • Attributes of visual perception
• Four umbrella principles of effective visualization
• Three-step process of executing an information display

Good Decisions Are Based On An Accurate Understanding Of Good Data

• 70% of the info we consume is visual. So, visual communications is very important.

• Visualization is the language of analysts.

• What are the types of data?

• Technique is more important than visualization tools.

• Trending charts (eg, line charts) don't make sense on categorical data.
(also, the intervals we make e.g., 0-10, 10-20, etc. also comes in categorical category)

• Take some time to contemplate on how you're gonna represent the data/result. It will save you
hours of unnecessary debate when you're presenting your result.

• What are the benefits of visual

representation?

• Job of an analyst: Business has to improve. They do analysis to help business improve. And,
they need to communicate it to the stakeholders so that right decisions can be made.

• Sometimes, the visual representation itself are so complicated that people spend more time
understanding the visuals than what it is trying to communicate.
 How do you help them focus on what is important, and how do you get them not
focus on what is not important.

• What are the four umbrella These conditions are necessary but not sufficient for optimal representation of data.
principles of effective visualization?

• When you present a chart, these are the four things that you have to absolutely follow.

• Write it down and go through this checklist when creating a visual.

• Describe the four umbrella

principles one by one:

1. Know your purpose

• A purpose is not necessarily a

message. How?

• Why does this representation has to be here?

• Once you have a purpose, it will automatically determine the form of the representation.

• Sometimes, a problem can't be solved by using just one visual representation.

In that case each chart or visual will serve as one step in the larger problem that you're trying to
solve.

• When I put a representative graphic together, the purpose of that should say this is going to
make my ability to communicate this complicated concept more easier.

----------------------------------------
• Umbrella principles means it covers everything.

• What is the purpose of having the handle that is shown?

- The handle emphasizes the concept of the umbrella principles. So, it is there for a reason.

• Literally, everything that's there, there has to be a reason.

So, if you look at it and say why is this there? What is the purpose of having that particular
graphic on the screen?
You should have a clear purpose.

-------------------------------------------------
• A purpose is not necessarily a message.

• A message could be communicated over a series of visual representations, and each of those will
have a purpose to contribute towards communicating that message.

2. Ensuring integrity
 • Typically you will have errors of omission and commission(?) when you're presenting data.

• Even a typo can render the integrity of the presentation questionable.

• Compare these two graphs (on

how the truth can be
distorted):

• It seems L is so volatile, but in reality it is not.

When we look 1999 to 2000 graph in L, one would say we grew 6x (we're very quick to measure
lengths), but's not the reality.

• L: The y-axis has been cut off.

• The integrity has been lost. The stories can be made biased hugely using these tactics.

3. Maximize data ink; minimize non-

data ink

• Spend ink on the item you want to show.

Do not spend ink on items that are not critical to the thought process.

• Any color ink that that does not serve a purpose and communicate no information to the end
user is non-data ink.

• Styling should be very specific for a purpose.

• Compare these two charts on

ink and non-data ink

• L: All the gridlines and the yellow colours are all non-data ink. It doesn't convey any additional
info.

• R: The title does communicate something. So it is data ink.

4. Annotations

Annotate to help the users.

• Compare these two graphs on
annotations.

• L: gives you an axis, and then you have to use a finger or a ruler across this to measure and read
off what those bars are.

• R: It's simpler to get rid of the axes tick marks and annotate the data. It visually looks so much
easier to read.
------------------------------------
• Is it necessary to annotate • Now, as you have more data points and the graphs get more crowded, annotation starts to look a
every single point? What if bit messy.
when the graphs get crowded?
• In that case, you do not have to annotate every single point. You annotate critical points.

• Executing information display is a

three-step process. What are those
three steps? Explain one-by-one.

Defining Message:

How do you ensure that when the person sees the graph, they get the same message that you're
trying to communicate?
○ Somebody is going to pull up that graph without the benefit of having you next to them to
explain that graph.
○ It's your responsibility to ensure that anybody who sees it should be able to understand
what it says.

It should be understood at a glance.

Choosing form:

Sometimes the best way to communicate is to just write a paragraph. There is nothing that says
that you have to show it graphically.

Choose a form of the visual display that is best aligned to the message you're trying to
communicate.

Creating Design:

Design principles come into play to ensure that visual cognitive cognition is easy.

Choosing the right form is easier because once you get a clear message, the form almost suggests
itself. But the design principle is what makes a difference.