Data Visualization Design Flashcards

Question 1

What is a visual encoding in data visualization?

Accepted Answer

A visual encoding is the mapping that connects a data value to a visual property, such as position, length, color, or size. Each mark on a chart carries information through these properties. A good design chooses encodings that match the kind of data being shown, so readers can decode values quickly and accurately.

Question 2

Which visual properties do people judge most accurately when comparing quantities?

Accepted Answer

Position along a common scale is judged most accurately, followed by length and direction. Angle and slope come next, while area, volume, and color saturation are judged less precisely. For this reason, quantitative comparisons should rely on position or length whenever possible.

Question 3

Why should a bar chart start its value axis at zero?

Accepted Answer

A bar chart should start at zero because bar length encodes the value, so a truncated axis exaggerates differences between bars. Readers compare bar lengths against the baseline, and a nonzero start makes small gaps look large. Line charts are less constrained because they encode values by position.

Question 4

What does the principle of data-ink ratio recommend?

Accepted Answer

The data-ink ratio principle recommends that most of a graphic's ink should present data rather than decoration. Designers remove or lighten gridlines, heavy borders, shading, and redundant labels that do not add information. The goal is a clean display where every mark earns its place and the data stands out clearly.

Question 5

What is preattentive processing, and why does it matter for charts?

Accepted Answer

Preattentive processing is the rapid, effortless detection of certain visual features, such as color hue, orientation, and size, before conscious attention engages. Designers use these features to direct the eye to important patterns or outliers in a display. Overusing them, however, can make a chart busy and reduce their power to draw attention.

Question 6

What is the difference between nominal, ordinal, and quantitative data?

Accepted Answer

Nominal data are categories with no inherent order, such as country names or product types. Ordinal data are categories with a meaningful order but unequal or unknown spacing, such as survey ratings from poor to excellent. Quantitative data are numbers with meaningful arithmetic, such as income or temperature.

Question 7

Why is a pie chart generally a weaker choice than a bar chart for comparing values?

Accepted Answer

Pie charts encode values through angle and area, which people judge less accurately than length or position. Comparing slices of similar size is especially difficult, and many slices make the chart hard to read. A bar chart places the same values on a common baseline, so differences are easier to see and rank.

Question 8

What is a common scale, and why does it help readers?

Accepted Answer

A common scale places several values along one shared axis with the same units and starting point. Because every value is read against the same reference, readers can compare them directly without mentally converting between separate scales. Small multiples that share a scale are a practical example of this principle.

Question 9

Which chart type best shows change in a value over time?

Accepted Answer

A line chart is usually the best choice for showing change over time, because the connected points emphasize the sequence and direction of the trend. Time runs along the horizontal axis, which matches how people naturally read temporal order. Bar charts can also work for a few periods, but lines are clearer for many.

Question 10

When is a scatter plot the most suitable chart?

Accepted Answer

A scatter plot suits situations where you want to examine the relationship between two quantitative variables. Each point represents one observation, positioned by its values on both axes. Scatter plots reveal correlation, clusters, gaps, and outliers, which summary bars or averages would often hide.

Question 11

Why are horizontal bar charts often preferred when category labels are long?

Accepted Answer

Horizontal bars give long category labels room to read naturally along the vertical axis, instead of being rotated or squeezed beneath short columns. The bars still share a common baseline, so comparison remains easy. Ranking the categories from largest to smallest then produces a clean, readable order.

Question 12

What is a histogram used for, and how does it differ from a bar chart?

Accepted Answer

A histogram shows the distribution of a single quantitative variable by grouping its values into adjacent intervals called bins. Bars touch because the bins form a continuous range, while bar charts separate bars to show distinct categories. Histograms help readers see shape, spread, and skew in the data.

Question 13

When would a heat map be a sensible choice for displaying data?

Accepted Answer

A heat map suits a dense grid of values where the pattern across two categorical dimensions matters more than exact readings, such as hourly activity across weekdays. Color intensity encodes each value in a cell. Because precise comparison of color is difficult, heat maps work best for spotting hotspots and broad trends.

Question 14

What makes a box plot useful, and what does its box represent?

Accepted Answer

A box plot summarizes a distribution using its median, quartiles, and typically its extremes or outliers. The box spans the middle half of the data, from the first quartile to the third quartile, with a line marking the median. Box plots let readers compare distributions across several groups in a compact space.

Question 15

Why might a stacked bar chart make comparison difficult for readers?

Accepted Answer

Only the bottom segment of a stacked bar sits on a common baseline, so its values are easy to compare. Middle and top segments float at varying starting points, making their lengths hard to judge. Readers can misread these segments unless the chart is simplified or the segments are shown side by side.

Question 16

How does a choropleth map differ from a symbol map?

Accepted Answer

A choropleth map shades predefined regions, such as counties or countries, according to a value for each region. A symbol map places markers such as circles at specific locations, with size or color showing values. Choropleths suit rates normalized by area or population, while symbols suit absolute counts at fixed locations.

Question 17

What are the three main types of color scales used in charts?

Accepted Answer

The three main color scale types are sequential, diverging, and qualitative. Sequential scales run from light to dark for ordered values from low to high. Diverging scales use two contrasting hues around a neutral midpoint for values that deviate above and below a meaningful center. Qualitative scales use distinct hues for unordered categories.

Question 18

Why is a rainbow color scale often discouraged for quantitative data?

Accepted Answer

A rainbow scale has no natural ordering of its hues, so readers cannot easily tell which colors are higher or lower. Its brightness also changes unevenly, which can create false visual boundaries where the data has none. Perceptually uniform sequential palettes generally communicate magnitude more faithfully.

Question 19

Why should you avoid relying on color alone to distinguish categories?

Accepted Answer

Some readers have color vision deficiencies and cannot reliably separate certain hues, so color-only encodings can exclude them. Pairing color with shape, labels, pattern, or line style makes the chart readable for everyone. Redundant encoding also helps when charts are printed in grayscale or viewed on dim screens.

Question 20

What does color contrast mean, and why does it matter for text and marks?

Accepted Answer

Color contrast is the difference in lightness between two colors, such as a label and its background or a data mark and the page. Low contrast makes elements hard to see, especially for people with low vision or when a chart is projected. Checking contrast helps ensure that labels and marks remain legible.

Question 21

When should a diverging color scale be used instead of a sequential one?

Accepted Answer

Use a diverging scale when values have a meaningful midpoint and you want to show whether data sit above or below it, such as temperature deviations from average. The two hues signal opposite directions, while lightness shows distance from the center. Sequential scales suit values that only increase from a minimum.

Question 22

Why should designers use restrained color by default in data graphics?

Accepted Answer

Restrained color keeps attention on the data and makes highlighted elements stand out. When every series is brightly colored, nothing in particular draws the eye. A common approach is to set most elements in neutral gray and reserve a strong hue for the single comparison or finding the chart is meant to communicate.

Question 23

What is a color legend, and what must it always make clear?

Accepted Answer

A color legend, sometimes called a key, explains which color corresponds to which category or value range. It must always make clear the units, the direction of the scale, and the meaning of each endpoint. Readers should not have to guess how to decode the colors.

Question 24

Why is it useful to test a palette in grayscale before publishing?

Accepted Answer

Converting a chart to grayscale reveals whether the colors differ enough in lightness to remain distinguishable when printed or viewed without color. Pairs of colors that look distinct in full color may collapse into similar gray tones. This quick check exposes weak palettes early, so you can adjust them before readers encounter the problem.

Question 25

Why should every chart have a descriptive title that states its message?

Accepted Answer

A descriptive title tells readers what the chart shows and, ideally, what it means. Titles that state the finding, such as sales rose during the second quarter, help readers grasp the point immediately. Without a clear title, readers must work out the purpose on their own, which slows interpretation.

Question 26

What information should a well-labeled axis include for readers?

Accepted Answer

A well-labeled axis shows the variable name along with its units, such as dollars or kilograms. Tick marks should use values that are easy to read, and the scale should start and end at sensible points. Clear axis labels remove guesswork about what each position or length represents.

Question 27

Why is direct labeling often better than a legend?

Accepted Answer

Direct labeling places each series name beside its line or bar, so readers do not have to move their eyes back and forth to decode colors. This reduces memory load and speeds up reading. Legends still serve well when many series or dense markings make direct labels crowded.

Question 28

What role do annotations play in a data visualization?

Accepted Answer

Annotations add short notes that point to specific features, such as a peak, a turning point, or an outlier. They guide readers toward the most important interpretation without requiring them to search the chart. Annotations should be brief and placed close to what they describe.

Question 29

How should a designer handle a reference line, such as an average?

Accepted Answer

A reference line should be labeled directly with its meaning and value, such as the average or a target. It should be visually subtle, often a thin dashed line in a neutral color, so it supports the data without competing with it. This makes comparisons with the benchmark quick and clear.

Question 30

Why is it important to show the source and sample size of data?

Accepted Answer

Readers need the source to judge how far to trust the data and whether it was collected in a reliable way. Sample size indicates how much weight a pattern can bear, since small samples produce unstable estimates. Stating both in a footnote lets readers evaluate the visual claim themselves.

Question 31

What is the purpose of a caption or subtitle in a chart?

Accepted Answer

A caption or subtitle adds context that the title cannot hold, such as the time period, geographic scope, or measurement method. It helps readers interpret the chart correctly without reading surrounding text. Captions are most useful when they are short and specific.

Question 32

How should text be sized and placed to stay readable within a chart?

Accepted Answer

Text should be large enough to read at the size the chart will be viewed, with consistent sizes for similar elements. Labels should sit close to the marks they describe and avoid overlapping lines, bars, or each other. Rotated text is harder to read, so horizontal labels are preferable when space allows.

Question 33

How should a designer decide whether to show raw counts or rates?

Accepted Answer

A designer should show rates or proportions when group sizes differ, because raw counts from larger groups naturally produce larger totals. Rates such as cases per thousand residents allow fair comparison across groups of different sizes. The choice should always match the question the chart is meant to answer.

Question 34

What is cherry picking in data presentation, and how does it mislead?

Accepted Answer

Cherry picking means selecting only the data points, time spans, or categories that support a preferred conclusion while omitting the rest. Such a chart can look convincing even when the broader data tell a different story. Honest designers show the full relevant range or explain clearly why a subset was chosen.

Question 35

Why should designers avoid 3D effects on bar and pie charts?

Accepted Answer

Three-dimensional effects distort perceived size, because perspective makes nearer or tilted elements look larger than they are. The extra depth also hides parts of the bars or slices, making accurate reading harder. Flat designs present the same values more faithfully and let readers compare them directly.

Question 36

How can a designer mislead readers through the choice of time period?

Accepted Answer

A designer can mislead by choosing a time window that starts at an unusually low or high point, which makes later changes appear more dramatic. Selecting a period that hides long-term context has the same effect. Showing a sufficiently long span, and stating the period explicitly, helps readers judge trends fairly.

Question 37

Why is it misleading to change the aspect ratio of a line chart to exaggerate a trend?

Accepted Answer

The aspect ratio controls how steeply a line appears to rise or fall, so compressing the horizontal axis can make a modest change look dramatic. Designers should choose proportions that reflect the data rather than the desired impression. Readers should be able to judge slope honestly from the chart alone.

Question 38

Why must the intervals on a value axis be equal?

Accepted Answer

Equal intervals are required because they keep axis distances proportional to data differences. Uneven tick spacing can make a steep slope look gentle or a small change look large, distorting the visual message. When a logarithmic or nonlinear scale is needed, it must be labeled clearly so readers do not misread it.

Question 39

Why should designers disclose data transformations such as smoothing or indexing?

Accepted Answer

Transformations change the numbers readers see, so a chart can differ from the raw values in important ways. Smoothing may hide spikes or sudden drops, while indexing rebases every series to a chosen starting point, which can make growth look different. Stating the method lets readers interpret the figure correctly.

Question 40

What steps can a designer take to check a chart for honesty before sharing it?

Accepted Answer

Check whether the axes, scales, and baselines match the question, and whether any omitted data would change the conclusion. Confirm that the units, sources, and sample sizes are stated, and that colors and labels do not exaggerate differences. Finally, ask a colleague to describe the main takeaway, then compare it with your intent.

Data Visualization Design Flashcards

1Perception and visual encoding

2Choosing chart types

3Color in data graphics

4Labels, titles, and annotation

5Honest and ethical representation

Learning Resources