STA191Data Visualization
Topic 1 · Introduction  ›  Week 1

Introduction to Data Visualization

The “why” week. Before we make a single chart, we learn the two ideas the whole course rests on: know your purpose and know your audience.

By the end of this week you can…

  • Explain what data visualization is and where it sits in the data-analysis process.
  • Tell exploratory from explanatory analysis, and pick which one a situation needs.
  • Ask the three context questions — audience, message, mechanism — before charting.
  • Describe how to meet a new dataset for the first time.
  • Justify why we visualize: to communicate, to turn data into information, and to show evidence.
Download this week: 📄 Notes (PDF) 📊 Slides (PPTX)

1.1  Why does a statistician care about pictures?

Statistics is the science of making decisions when data varies. But here is the catch: a correct answer that nobody understands changes nothing. A perfect analysis on a slide the audience can't read is a wasted analysis. Data visualization closes that gap.

Here is why it works. When you read numbers, your brain processes them slowly, one at a time, using the same system you use for words. But position, length and colour are processed by your visual system all at once — before you even consciously try. A chart quietly moves the hard work of comparing from slow reading to fast seeing. That is why the answer to "which month is highest?" arrives from a chart before you finish the question, but takes real effort to dig out of a table.

● Key idea

A table of 500 rows is data. A chart that shows those rows leaning one way is understanding. Visualization is a thinking tool you use throughout — first to help you understand the data, later to help others understand your conclusion. It is not decoration added at the end.

🔎 See it for yourself

Question: which month had the highest sales?

MonthJanFebMarAprMayJunJulAugSepOctNovDec
Sales (RM ’000)423847555160586349716862

Scanning the table takes a few seconds. Now press the button — the answer jumps out before you finish reading.

1.2  Two jobs, two mindsets

Every visual you make does one of two jobs. Naming which one you are doing is one of the most useful habits this course can build.

ExploratoryExplanatory
AudienceYou (the analyst)Others (decision-makers, public)
GoalDiscover what’s interestingCommunicate one clear message
How many chartsMany, rough, disposableFew, polished, deliberate
Mindset“What’s in here?”“Here’s what matters — and why.”

Exploratory is a conversation between you and the data — sifting through 100 rough stones to find one or two gems. Explanatory is a conversation between the data and the audience — showing those gems clearly.

Where this happens in STA191: the exploratory stage is what you'll do in the tool weeks (Google Sheets and Power BI, Weeks 5–11) — trying many quick charts. The explanatory stage is what Chapters 2, 3 and 5 are about — designing the few clean visuals and the report that go to your audience. Almost always you do exploratory first, then explanatory: exploration finds the message; explanation delivers it. They are stages, not alternatives.

● Example · Grab drivers in KL

A ride-hailing analyst has a year of trip data. Exploratory: she makes dozens of quick charts by hour, day, weather, district — and spots that demand spikes on rainy evenings in the city centre. Explanatory: for the operations manager she makes one clean chart titled “Rainy evenings drive 40% more trips in KL city centre — put more drivers on standby.” Same data, two different jobs.

● Watch out

Exploratory and explanatory are stages, not alternatives — you almost always explore first, then explain. And mind the spelling: explOratory (explore) vs explAnatory (explain).

1.3  Before you draw anything: three questions

The most common beginner mistake is opening a charting tool first. Stop and answer three questions:

  1. Who are we communicating to?
  2. What do we want them to know or do? (One sentence — if you can’t say it, you’re not ready to chart.)
  3. How can the data support that point?
● Example · One finding, three audiences

Finding: dengue cases in Selangor rose sharply this quarter.

  • State Health Director: detailed line chart, weekly cases by district — she needs precision.
  • Public on Instagram: one big number — “Dengue up 3× — clear stagnant water now.”
  • Fellow researchers: full time series with methodology and past-year comparison.

The data never changed. The purpose did.

● Try it · Who’s it for?

Take this finding: “UiTM students who attend tutorials score 12 marks higher on average.” Sketch — on paper — how you’d show it to (a) the Dean, (b) first-year students, (c) parents at open day. What did you emphasise differently each time?

● Session 1 recap

Visualization turns slow reading into fast seeing. Exploratory analysis (you + data) finds the gem; explanatory analysis (data + audience) shows it — explore first, then explain. Before charting anything, answer who, what and how: the audience, not the data, drives the design.

1.4  Meeting a new dataset

Before visualization comes understanding what you’re holding. When a dataset lands on your desk, resist the urge to chart it. First, interrogate it:

Then prepare it: check for missing values and outliers, rename confusing variables, and compute descriptive statistics (mean, median, quartiles, spread, skewness) to feel the shape before you plot.

● Example · Naming attributes

In a UiTM student dataset each row is one student, with attributes: Student ID, Programme, Semester, State, CGPA, Tutorial attendance %. In a kopitiam daily-sales dataset each row is one day: Date, Day, Total sales (RM), Customers, Weather, Public holiday? Before charting, what would you check? (A missing day? A CGPA of 4.5 that can’t exist? A state typed three different ways?)

● Connect it back

You already compute mean, median, quartiles and skewness. Visualization is just seeing them: a boxplot is a five-number summary you can look at; a histogram is a frequency table you can see.

1.5  What the experts say it is

The field’s leading authors converge on one idea — visualization is storytelling with data:

The common thread: clarity in service of a message — not “make it pretty,” make it clear.

1.6  Why we visualize — three reasons

1 · To communicate

Shape reveals patterns that stay hidden in a table — dots show relationships, lines show trends, bars show comparisons.

2 · To inform

We climb the ladder: data → information → knowledge → action. A chart that drives a decision has done real work.

3 · To show evidence

A graphic is visual proof the audience can see for themselves — but only as good as the message behind it.

How the shape carries the message
Dots → relationship
Lines → trend
Bars → comparison

The way you draw data decides what the eye notices: dots reveal a relationship, a line reveals a trend, bars reveal a comparison.

● Example · The ladder in action

Data: a clinic’s monthly patient counts. Information: a line chart shows visits spike every haze season. Knowledge: respiratory cases drive it. Action: the clinic pre-orders inhalers and adds staff before the haze arrives.

● Session 2 recap

Meet a dataset before you chart it — know its source, clean it, and read its descriptive statistics. Then visualize for three reasons: to communicate, to turn data into information, and to show evidence. A chart is only as good as the message it carries.

Key terms & definitions

Exploratory data analysis
Looking at data from many angles to discover what's interesting — for yourself. Many rough, disposable charts.
Explanatory data analysis
Communicating one specific finding to a specific audience. Few, polished, deliberate visuals.
Context
The three questions before charting: who is the audience, what is the message, how does the data support it.
Attribute
A characteristic that describes each record in a dataset — usually a column (e.g. CGPA, State, Sales).
Descriptive statistics
Summary numbers that describe a dataset's shape: mean, median, quartiles, spread, skewness.
Big Data
Data too large or fast to grasp unaided — not just "a lot" of data; visualization helps make sense of it.
Data → information → knowledge → action
The ladder a good chart climbs: raw numbers become an insight that drives a decision.

✅ Quick self-check

1. You make 30 rough charts to figure out what’s interesting in a dataset. Which job is this?

2. Which is not one of the three context questions to ask before charting?

3. “A clean upward line proves sales are recovering.” This is visualization used to…

Week 2 →