Visual Design, Clutter, and Pre-Attentive Attributes

This session goes one level deeper than the introduction: past the question of why visualization matters and into the question of how people actually read a screen. The plan is to recap the exploratory-versus-explanatory split and walk through how to choose an effective visual. The rest of the time goes to two ideas that decide whether a chart teaches or tires. One is clutter — everything that should not be on the page. The other is pre-attentive attributes — the visual properties our brain registers before we consciously look.

Everything in this session is tool-independent. The same rules apply whether you build the chart in Tableau, Power BI, Google Data Studio, or Qlik, and they apply just as much to a plain slide deck you present to stakeholders. Learning the concepts first, before any tool, keeps you from being biased by whichever product you happen to touch first — that is a deliberate choice the course makes on purpose.

2.1 Exploratory vs Explanatory Analysis

2.1.1 Two Modes, Two Audiences

Exploratory analysis is what you do to uncover patterns. Think of it as a hidden trail — you do not yet know what you are looking for, so you search. It is often interactive, because you are trying combinations and permutations: filtering here, zooming there, changing the grouping to see what falls out. The typical charts of the exploratory mode are the scatter plot, the histogram, and the box plot. The audience for exploratory work is usually an analyst or a small group of people who are part of the search.

The reference text describes exploratory analysis as hunting for pearls in oysters: you might have to open a hundred oysters (test a hundred hypotheses, or look at the data a hundred different ways) to find perhaps two pearls. The search itself is the point of the exercise — nothing is wasted, because every wrong view still teaches you something about the data.

Explanatory analysis is what you do when the outcome is finalized. It confirms something and tells a crisp story: this is what we found, this is what has happened. The typical charts are bar graphs, line charts, and heat maps. The audience is wider — the people who need the result, not the journey. The same reference text puts it sharply: when you move into explanatory mode, you stop showing all one hundred oysters and concentrate on the two pearls — the information your audience actually needs. A common, understandable mistake is to present all of your exploratory work as evidence of how much you did. Resist that urge; it forces your audience to reopen every oyster.

Two questions settle which mode you are in at any moment:

1. Who is the audience? An analyst or a small team searching alongside you, or a wider group that needs the conclusion?
2. What does the outcome look like? An open search that may not end anywhere yet, or a finalized result with a clear takeaway?

When they are answered the same way for the same visual, you are in one mode — and the chart choice, the amount of annotation, and the level of explanation all follow from that answer.

When to pick which: explore privately (or with your team) to find the story; explain publicly with a polished visual once the story is fixed. Do not present exploration as explanation.

2.1.2 Mini Quiz: Which Mode Is It?

The session used three true/false statements to check that the recap had landed. The quizzes are engagement tools — there is no right-or-wrong penalty, and everyone is encouraged to try. The answers, with the reasoning from the recap:

The three statements are the same idea tested from three angles: exploration needs no preset question, explanation is about the final outcome, and exploration guarantees nothing. If you can answer all three without hesitation, you have the split.

2.2 Choosing Effective Visuals

2.2.1 Start With the Goal and the Audience

Before picking any chart, be clear about two things: what you want to convey and who your audience is. The same data looks very different for students than for executives or stakeholders, so tailor the outcome to the reader. There is no hard-and-fast rule that a chart can never be swapped for another. Some charts work interchangeably. Still, certain chart types fit certain jobs better, and the preference comes from the goal and the audience first.

The reference text frames the same idea as three questions asked in order: who are you communicating to, what do you need them to know or do, and only then how — which data and which visual will make the point. If you cannot answer the first two concisely, no chart choice will save the message. A fourth consideration is the mechanism: a live presentation lets you control what the audience sees and when, so slides can stay sparse; a written report has to carry the explanation by itself, so it needs more explicit detail. The chart type you choose can be the same in both — but how much you spell out changes.

2.2.2 Chart Types and When to Use Them

• Line charts are preferred for a trend over time. Usually the time goes on the X axis and the number you care about — sales, traffic — goes on the Y axis. Because the points are physically connected, a line implies a continuous flow, which suits time series but not categorical data. One caution from the reference text: if you plot time, the intervals must be consistent — mixing decades with individual years makes the distances between points lie about the data.
• Bar charts are for comparing values between categories: product A sales versus product B sales, or how your customers are doing against each other. Bars are easy to read because the eyes compare the end points of the bars, all aligned on a common baseline. That baseline is the catch: bar charts must start at zero. The classic failure is a chart that starts the Y axis at 34 to dramatize a rise from 35% to 39.6% — the visual increase becomes 460% instead of the true 13%, and one discerning reader noticing it destroys your credibility. The same rule does not apply to line charts (they show relative position, not length from a baseline), but use a nonzero baseline for lines with caution and say it out loud.
• Pie charts need care. Use them only when you have little data and when you are showing a part-to-whole relationship — say 25 percent or 30 percent of the overall picture. The reason for caution: the mind cannot read diagonal lines and diagonal information, so pies are comparatively hard for the brain to process. The reference text goes further and calls pie charts evil — they ask the audience to compare angles and areas, which the human eye does poorly; when two slices are close in size you cannot tell which is bigger, and even when they are not, you cannot judge by how much. That does not make pies bad; it makes them a tool to use deliberately. When you can, a horizontal bar chart does the same job better: it keeps the part-to-whole idea (bars summing to the total) while giving the eyes a clean baseline.
• Scatter plots are for correlation: price versus demand, studying hours versus scores. They reveal relationship details well, because each point carries two values at once — one on the X axis, one on the Y axis. The reference text uses a bus fleet example: miles driven per month on the X axis, cost per mile on the Y axis, which shows cost per mile climbing when the fleet is underused or overdriven.
• Maps are the no-brainer for geographical data.
• Histograms are for showing distribution — how the data is spread across value ranges.
• Area charts are for cumulative values — the curve accumulates over time.

Real-world: this guidance is why you see line charts in finance dashboards, bar charts in media reporting, and maps in medical and government reporting. Schools use visualization, doctors and medical staff use it daily, and governments keep moving into digital dashboards. The reach of visualization keeps expanding, which is why getting the chart choice right matters — a finance team watching a trend needs the line, a media outlet comparing candidates needs the bars, and the choice in each case starts from the job, not from the tool's default menu.

2.2.3 A Quick Tableau Tour

The session paused the concepts to show how these chart types look inside Tableau, using the sample sheets the tool ships with. The point was not to teach the tool yet — dimensions and measures, and how many of each a chart expects, come in the dedicated Tableau class. The goal here was simply to see the options.

• Maps: drop in a country, states, or coordinates and the tool is intelligent enough to recognize geographical data and plot the map on its own. Power BI does the same.
• Area charts: good for cumulative stories, showing both the running total and what sits underneath.
• Heat maps: colors scale automatically; you can set min, max, or a range.
• Scatter plots for relationship mapping, bar graphs for comparison, and tooltips you can configure yourself.
• Text tables (cross-tabs) for showing text, highlight tables that auto-highlight, symbol maps, horizontal bars, stacked bars, circles, and line graphs — the classic for anything with a time series. Tree maps build the hierarchy automatically when the data has parent and child levels.

The neat behavior to remember: as you drag fields onto a sheet, Tableau automatically enables the chart types that qualify for that data, and disables the rest. The enabled options appear on the right-hand panel, where you can also add labels and colors. The tool literally tells you what is possible with the data you have placed.

2.2.4 The Chart Cheat Sheet

A second view grouped charts by the job they do, and it is a good reference to keep:

• Inform — when you just need to state a single fact, use one big number, a donut chart, or a pictograph. "25 percent of something" reads much clearer as a single visual than inside a table. The reference text agrees: when you have one number or two to share, simple text — the number made as prominent as possible with a few supporting words — beats a graph that makes the audience hunt for the digits.
• Compare — two or more values: bar graph, bubble chart, and similar options.
• Change over time — line graph, area chart, timeline, map chart.
• Organize — Venn diagrams, mind maps, tables.
• Relationship — scatter plot for how one variable changes with another, histogram for how your data is distributed, multi-series charts.

Two practical notes from the discussion. First, not every chart exists in every tool — these are concepts, not menu items, and the same thinking applies in your PPTs, in Google Studio, wherever possible. Second, the pie limit again: past 10 to 15 data points the slices blur and the percentages become indistinguishable. Pick and choose intelligently based on your data.

2.3 Design Principles for Visuals

2.3.1 Keep It Simple

A set of design habits came before the formal clutter discussion. Keep the visual simple: use clear labels, titles, and legends, and highlight whatever you want your user to see. Do not try to put so much text on the page that everything points at everything else. We often feel obliged to fill the sheet — we should not. Ask what exactly the aim of the page is and let that decide what stays.

The reference text adds a useful test from product design: does this element afford the right reading — that is, does it make obvious how the page should be used? A title that names the takeaway, an axis that is labeled, a legend that sits near its data — these are the affordances that let good design fade into the background. When every element competes, the design becomes the noise.

2.3.2 Use Color With Purpose

Color cannot be picked at random. Our mind is wired so that red usually means negative and green means positive — you have seen this convention in many data products. The demo showed how it works in practice. With a gain-loss field, you drag the field onto the color shelf and the tool assigns a scale automatically, complete with min and max. You can override the range. Choose a red-green diverging palette and positive versus negative becomes visible instantly. The point of the demo: color has to be chosen for a purpose, not because a color is bright. Shapes and sizes also represent value, and whenever you position an element, position it on a coordinate basis.

Three refinements worth carrying from the reference text. First, use color sparingly: the more colors you add, the less any one of them stands out — a palette of many hues loses its signal. Second, use color consistently: the same meaning should wear the same color everywhere in a deck, because the audience learns the mapping once and then assumes it holds. Third, design with the colorblind in mind — roughly 8% of men and about half a percent of women have color-vision deficiency, most commonly red-green — so pair the red-green convention with a second cue (bold, or a plus/minus sign) rather than relying on the hue alone.

2.3.3 Make It Interactive

Interactivity is not a luxury. Gone are the days when users look at static data — interactive visuals help you engage business users. Give filters and let users choose what they want to see: for multi-year data, give a filter on the year; for many products, give a product filter. The filter also reduces the content you are forced to keep in one view — you no longer need to cram everything onto a single page. Users feel more engaged and more curious, and that keeps them in the data.

2.3.4 Iterate With Your Audience

You cannot make a good visualization in one pass. You show something, the user says "this is not clear, I am not sure how this data links to that," and you adjust. Keep showing the visual to end users, keep taking feedback, keep enhancing it. Move data to another sheet when the user says it makes more sense there. Changing your visualization is not a bad thing — the ultimate aim is that your users actually understand.

The reference text offers a cheap way to test any visual before it ships: show it to a colleague and ask them to talk through where their eyes go first, what they see, what observations they make, and what questions they have. That one conversation exposes most of the assumptions a designer cannot see in their own work.

2.3.5 Stay Consistent and Accessible

Visual consistency matters across a set of sheets or a deck. One slide cannot use blue for one thing and the next slide use blue for something else. Use a color for one meaning and keep it that color everywhere so people can relate easily. Also consider users with different abilities — not everybody sees things normally. Add descriptions, and remember keyboard navigation exists so users are not forced to rely on a mouse. These may not be used every day, but they belong in a well-built visualization.

The reference text frames this as accessibility: design usable by people of widely varying technical skill and ability. The onus is on the designer — it should not take an engineering degree to understand an engineer's chart. Two concrete moves: do not overcomplicate (a visual that looks hard to read is read less), and treat text as your friend (every chart needs a title, every axis needs a title, and if there is a conclusion you want the audience to reach, state it in words).

2.3.6 Build a Logical Flow

Think logically and lay out a sequence the user can follow. You cannot present the end state, then jump back, then forward again — the reader should not have to go zigzag through your story. Sequential information is what lets the user relate one part to the next.

2.4 Clutter

2.4.1 What Clutter Is

Clutter is everything unorganized and unnecessary that the brain struggles to process. People say "your desk is very cluttered" — a desk covered with unorganized things — and the same phrase describes a bad visual. Clutter is one of those problems that usually goes unnoticed. We do not realize how difficult we are making it for the end user to process the data. No wonder that within five or ten minutes they start looking elsewhere. We are burdening the brain and the eyes of the very people we are trying to reach.

The goal of a visualization is simple: look at the screen, look at the slide, and take a decision. The visual should tell you that here sales are going down and here sales are going up. When it cannot, the mind gets confused — should I look at the top of the screen or the bottom? Here something is underlined, there something is bold, and all of it has meaning for the brain. We cannot place everything on the page. When we do, the data is there, but the eyes cannot reach it and the point is lost.

The mechanism: grids, labels, legends, unnecessary shadowing, and background lines. The brain visits each element one by one and decides whether it is useful. The more content you put on the page, the more stops the brain makes. After a few seconds or a few minutes the brain says: I give up, I cannot process this. Titles are required — the title is one of the most important parts of any graph — but color combinations, positioning, and fonts decide whether the page reads easily.

A useful lens for telling signal from noise comes from the Gestalt principles of visual perception — the rules, studied since the early 1900s, that describe how the brain groups what it sees. Six matter here:

• Proximity — objects close together read as one group, so spacing alone can organize a page.
• Similarity — objects sharing color, shape, or size read as related.
• Enclosure — objects inside a box or shaded region read as one group.
• Closure — the brain completes an incomplete figure, so a chart still reads as a whole even without a border around it.
• Continuity — the eye seeks the smoothest path, so elements that line up are seen as connected.
• Connection — physically connected objects read as related, which is why lines connect the dots in a line chart.

Gestalt tells you why a border feels unnecessary (closure: the chart is a shape on its own), why spacing can replace dividers (proximity), and why a shaded region separates forecast from actual (enclosure). Anything that fights these grouping rules tends to be the clutter your audience pays for.

2.4.2 Cognitive Load

The word for all of this processing is cognitive load — how much processing the brain has to do.

Clutter makes the CPU choke. The consequences are concrete. The user is distracted and struggles to find the purpose of the page. The extra data invites misinterpretation, because anyone can read the jumble however they want. Trust goes — the user starts wondering what is being said and which one is right — and confidence in the report erodes. This is how presenters unknowingly let their audience drift away.

2.4.3 Worked Example 1: One Chart, Cleaned Up

The first example shows a chart before and after a careful cleanup — the same data, transformed by small changes.

White space played its part too. We often feel that leaving blank space is a bad thing — wherever there is a space, put something, the user will read it. The opposite is true: the more text you put, the more the user runs away. Blank space is like letting the user breathe. Margins are given conveniently, the eye has room, and the reader has time to process instead of the mind going into overload. The reference text compares white space to pauses in public speaking: a well-placed pause lands a statement, and a page that never pauses suffocates the audience. Use white space strategically — if one number is the whole message, make it the only thing on the page.

2.4.4 Worked Example 2: Shapes vs a Clean Bar Chart

The second example is the same data shown two ways.

2.4.5 Worked Example 3: Tickets Received vs Processed

The third example is a line chart for an organization that tracks support tickets raised month by month. The blue line is the tickets received; the red line is the tickets being processed. It is a simple requirement, and eight out of ten people would have drawn the graph the way the original was drawn. There is nothing wrong with it, and the trends are visible. The exercise is to see how small tweaks make it better, still using the same line chart.

2.4.6 How to Avoid Clutter

The checklist from the session, worth keeping as a habit:

1. Keep it simple — the fewer elements, the fewer decisions the brain makes.
2. Eliminate — only essential things belong on the page. You do not need to show a ten-year history at once.
3. Give filters — wherever possible, let the user filter and digest the data one slice at a time. The filter is what saves you from cramming every year and every product into one view.
4. Use visual hierarchy carefully — placement, sizes, and colors must be consistent; you cannot use different treatments on different slides; it all has to flow.
5. Less is more — make white space your friend. There is no harm in leaving a little of the page free for the user to digest.

The reference text adds one discipline to the second item: when in doubt about cutting an element, ask "would eliminating this change anything?" — if the answer is no, take it out. And push necessary-but-secondary items to the background (light grey) rather than removing them entirely. One caution on the other side: not every extra mark is clutter. Currency signs, percent signs, and commas in large numbers are not decorative — they are context the reader needs, and stripping them makes numbers harder to interpret, not easier.

These concepts are not tied to any tool, so do not worry about how they will happen in Tableau or any other product. They are the things you keep in mind when you place controls on a form or build a page. And they arrive gradually: today you might learn that an X axis should be horizontal instead of diagonal, and that is one more habit in your pocket. Nobody becomes a visualization expert overnight; mastery comes with time and application.

2.4.7 In Conversation: Clutter at Work

Two exchanges from the discussion period, when the class was asked whether any of this matched what they see at work.

The second student shared a habit many presenters already have without naming it — mentally testing the visual from the audience's side.

2.5 Pre-Attentive Attributes

2.5.1 What Pre-Attentive Attributes Are

Pre-attentive attributes are the natural extension of the clutter discussion: they explain a little more about how the mind processes and sees things without us realizing. A pre-attentive attribute is a visual property the brain registers before attention gets involved — the processing happens in a fraction of a second. No training is required. Our mind and brain are naturally tuned to notice these properties, which is exactly why they are so powerful: they will draw attention whether we plan for it or not. That is why we should use them wisely — and why, if we do not want something to receive unnecessary attention, we should avoid giving it one of these attributes.

The reference text shows the same effect with a block of numbers: counting the 3s with no visual cue is a hunt through four lines of digits; repeat the exercise with the 3s rendered in a distinct color intensity and the answer is immediate — you do not have time to think, and suddenly there are six 3s in front of you. That is iconic memory being used: the brain picks out the difference before you know you are looking for it. This is why a well-designed visual can let the audience see what you want them to see before they know they are seeing it.

2.5.2 The Attribute List

• Length. In the sorted bar chart seen earlier, the eye relates to the data easily: high, then lower, then lower. Rather than reading the numbers, the length of the lines lets the eye grasp the scale and visualize the size of the values.
• Line width. Making one indicator bold lets you differentiate it — less widely used, but handy for emphasis.
• Size. From a consistent font, one element elevated in size catches the eye immediately. This is the "one big number" effect.
• Curvature. A subtler one; usage varies with the use case.
• Enclosure. Whenever you enclose something, the eyes go to it — the brain asks "why is there a box here?" and looks.
• Color intensity. Things you do not want the user to read carefully can be pushed to gray; things you want read carefully can be made black. One information recedes into the background, another comes to the foreground.
• Orientation. A tilted element draws attention — which is why a 45-degree legend is a mistake: it forces the reader to tilt along with it.
• Shape. If everything is a circle and one item is a diamond, the unique shape is caught instantly.
• Position. Where you place a legend, a call-out, an element — position steers where the eye goes.
• Animation. Moving elements get more attention from the brain; animated arrows and motion are used exactly for this.

One important property of the list: some attributes carry quantity and some only carry category. Most people read a longer line as a larger value — that is why bar charts are easy — and size, line width, and color intensity can suggest relative value. But hue cannot: "which is greater, red or blue?" is not a meaningful question. Use the quantitative attributes (length, position, size) to encode how much, and the categorical ones (hue, shape) to encode which group something belongs to.

2.5.3 Why They Matter

Pre-attentive attributes do four jobs in a visualization. They guide attention: you can point the user at exactly where you want them to look. They help prioritize: make something bold, use contrast, and the eye catches it first. They reduce cognitive load: instead of scanning everything, the user is told "look at that data, not elsewhere," so the working memory is spared. And they make the visual more readable and more engaging. Used together, they automatically create a hierarchy: the user sees the top item, then looks left, then right. You can navigate the end user's attention and control the order in which information flows.

There is a time budget behind these four jobs: studies put the initial window of attention at about three to eight seconds — during that window the audience decides whether to keep looking or move on. If the pre-attentive attributes have been used well, even that short window delivers the gist of the message. If they have been used badly — everything different, nothing prioritized — the audience spends the window lost.

One caution from the reference text: highlighting one aspect of a visual can make the rest harder to see. In exploratory analysis, where you are still looking for the story, it is usually better to hold back on pre-attentive attributes; in explanatory analysis, where the story is fixed, they are exactly the tool for making that story visually clear.

2.5.4 Cognitive Load, Defined

The formal definition lands here. Cognitive load is the amount of information our working memory can handle. The comparison from the session: computers have processing power and RAM, measured these days in gigabytes and terabytes. The brain also has to do work; whatever the working memory can handle is the cognitive load. The more cognitive load a visualization demands, the sooner the user struggles, and after a few seconds information starts slipping from their attention. Keeping the cognitive load in the right range is exactly what the clutter removal and the pre-attentive attributes achieve together — clutter takes load away by deleting noise, and attributes take load away by telling the brain where to look instead of making it scan everything.

2.5.5 How and Where to Use Each Attribute

• Size represents higher values. In a slide deck you will see one number made very big, and that number is the message: look at this one, the others are different.
• Color highlights the important elements — make one color bright or contrasted and the eye differentiates it instantly without scanning the rest.
• Positioning matters for every element: where you place your legend and your call-outs steers the reading order.
• Orientation draws attention when used deliberately; do not make your user tilt their head at a slanted legend while other information competes.
• Shapes elevate one item out of the information — a unique shape is caught immediately.
• Animation — arrows, motion, animated charts — gets more attention from the brain.

2.5.6 Worked Examples: Reading Charts Through Attributes

A series of before-and-after charts made the attributes concrete.

In Tableau all of this is a matter of commands: select a single bar and recolor it, gray out the rest, or color-code only the top three. The attributes get applied based on which data you want to elevate.

2.5.7 A Word of Caution

Two warnings to carry forward.

Learn the concepts, then apply one improvement at a time — your colleagues will start noticing your data visualization getting better before you even realize it is happening.

2.6 Design Concepts and Well-Built Examples

2.6.1 Storytelling

Storytelling is a design concept that helps make things simpler — data flows through the visual like a narrative. You can build a dashboard that runs like a slide deck. It shows the first output, then the second, in a flow you control, combining sheets and dashboards into one storyline. Storytelling is part of the course syllabus — it comes up in the Tableau part of the course, where stories are built from the visuals. Stories also build confidence: when the data tells a coherent story, the audience trusts it.

The reference text treats story as the point of the whole exercise: good data visualization is not a collection of facts on a topic, it is a narrative. Facts become information only when they are arranged into a sequence that leads somewhere — and that arrangement is the designer's job, not the chart's.

2.6.2 Visual Hierarchy

Visual hierarchy is how you prioritize your information — deciding which thing your user sees first.

2.6.3 Less Is More, Consistent and Cohesive

The session closed the loop on the recurring design lines: less is more, white space is your friend, and the visual must be consistent and cohesive across pages. One line is worth framing: use your color as a language. You cannot pick colors randomly and expect them to work. Color carries meaning, so choose it to match your storyline and use contrast deliberately. If blue means "our products" in one chart, blue means "our products" in every chart; if grey means "not the story," grey stays background material everywhere.

2.6.4 Interactivity in Practice

A short live demo showed how cheap interactivity can be in Tableau. There is a filter pane; you drag a field onto it, choose "show as filter," and the filter appears — that is all. Filters can be of different types — multi-select, radio buttons — and a filter on one visual can drive another. The same is true in Power BI, and the effect is real: interaction adds a dimension that a static page cannot have, and lets the user play with the data.

2.6.5 Worked Example: A Fundraising Campaign Dashboard

A well-built dashboard for a fundraising campaign shows how far deliberate attribute choices go.

2.6.6 Worked Example: Sales Actuals vs Forecast

A sales-over-time chart shows how to separate reality from expectation.

2.6.7 Worked Example: Missed Targets in Red

A project-tracking visual shows how color and commentary cooperate to tell one chosen story.

The takeaway, repeated in the session: no graph is right or wrong, good or bad. It is the way we select, present, and let others interpret the data that makes the difference.

2.7 Appendix Concepts: Cognitive Load and Data Ink Ratio

2.7.1 Cognitive Load

Cognitive load is the amount of information the brain can hold at once — the working-memory budget.

2.7.2 Data Ink Ratio

Data ink ratio is a measure of how much ink (or pixels) a visual spends to present its data. The description from the discussion: how much data we are able to present using that ink or pixel. The idea behind it:

2.8 Course Roadmap and Learning Resources

2.8.1 How the Course Flows

The first classes of the course are conceptual on purpose. These design concepts are not tied to any tool. Learning them early keeps you from being biased by whichever product you first touch. Once you are inside a tool, learning gets focused and limited to what that tool offers. Next comes the taxonomy of charts and plots, spread over the next couple of classes. The last class of this module compares the available tools — Power BI, Tableau, Google Data Studio, and a tool referred to as "Click" (likely Qlik). After that, two modules go deep into Tableau with hands-on practice, and then a module works with Python and its visualization libraries. The Tableau flavor in the early classes is a taste of what is coming — the share of tool work grows as the course progresses.

The order is deliberate. Concepts first, tools later — so that when the Tableau classes arrive, the chart choices are already habits, and the tool is only a matter of where the buttons are. Nothing tool-specific (dimensions, measures, filter mechanics) is expected before those classes.

2.8.2 Learning Resources

The example charts shown in this session come from a textbook the course refers to as textbook number two. The examples appear there, though the concepts exist in many places and no single book is required. There is also a YouTube video made by the Tableau team itself about designing in Tableau. It is useful for seeing these concepts applied to shapes, sizes, and colors inside the tool.

Two ways to use the video: watch it once to see the concepts applied in a real product, then revisit specific minutes as the Tableau classes arrive — the design concepts it shows are exactly the ones from this session, only executed with the tool's menus.

Exam Guidance Summary

No formal exam details were given in this session; what follows is the honest summary of what was signaled as important and how the course is structured.

• Quizzes are engagement tools. The three true/false questions on exploratory vs explanatory analysis carried no penalty. Everyone was encouraged to try, and the answers were confirmed in class. The answers: no predefined research question is required for exploration; explanatory focuses on why; exploration gives no definite answers. Expect these low-stakes checks to keep appearing.
• Chart selection is core material. Know which chart fits which job: line for trends over time, bar for comparison between categories, pie only for small part-to-whole relationships. Do not push pie past 10-15 data points. Also know scatter for correlation, map for geography, histogram for distribution, area for cumulative values.
• Clutter and pre-attentive attributes are the heart of this module. Be ready to identify clutter in a chart: borders, grids, tilted legends, excess labels, missing white space. Also explain cognitive load and its costs: misinterpretation, lost trust, users drifting away. Name the pre-attentive attributes and say what each does to the eye: length, line width, size, curvature, enclosure, color intensity, orientation, shape, position, animation.
• Study advice. Read the appendix concepts offline (cognitive load, data ink ratio). Watch the Tableau team's YouTube video on designing in Tableau. Apply one improvement at a time — the course itself says mastery is gradual, not overnight, and that these concepts apply to any presentation, not only to dedicated tools.
• Course structure worth remembering. Taxonomy comes next, then a tools comparison around the fifth class (Power BI, Tableau, Google Data Studio), then hands-on Tableau modules, then Python. Nothing tool-specific (dimensions, measures, filter mechanics) is expected before the Tableau classes arrive.

How to revise this session: first, check yourself against the three quiz statements from section 2.1 — they are the fastest test of the exploratory/explanatory split. Second, reproduce the cheat sheet from memory (inform, compare, change over time, organize, relationship) and attach one example chart to each job. Third, walk through the tickets chart from section 2.4 and name every element you would remove and why — that single exercise covers clutter, cognitive load, and the pre-attentive lesson of direct labels in one pass.

Key Industry Applications

• Classic visualization industries (from the recap): finance, media, medical — the fields where visualization first became critical. Finance dashboards lead with line charts for trends; media reporting leans on bar charts for comparisons; medical reporting uses maps and geographic views.
• Everyday and government use: schools and teachers, doctors and medical staff, and governments moving into digital dashboards.
• Project management and sprint boards: tools like sprint boards produce bar and line graphs out of the box. The presenter's real job is choosing which graph serves the meeting — weekly status vs monthly review.
• Support operations: teams track tickets received vs tickets processed to watch backlog gaps grow (the July-to-August widening in the worked example). The cleaned chart turns that tracking into a hiring or staffing argument.
• Fundraising: campaign dashboards showing progress to date (33,000 raised) against the goal and last year's results (500,000) on a single view — one bold number carries the whole story.
• Sales planning: actuals vs forecast charts with solid/dotted lines and shaded forecast regions for instant reading of where fact ends and prediction begins.
• Project portfolio reviews: missed targets flagged in red (42 percent missed, one third of projects) so the problem areas surface first, with commentary matching the color.
• Any stakeholder deck: the same design concepts apply whether the output is a PPT, Power BI, Tableau, Google Data Studio, or Qlik — the concepts come first, the tool is just where the buttons are.

The pattern across all of these: every one of them is the same session applied to a different domain — decide the goal and audience, pick the chart that fits the job, remove the clutter, and use the pre-attentive levers to say the one thing the page is for.