Common Graph Types

Why This Matters

Data visualization is everywhere—from news articles and scientific studies to business reports and social media infographics. On your exam, you're being tested on more than just recognizing what a graph looks like. You need to understand when to use each type, what relationships it reveals, and why one graph might be better than another for a given dataset. These skills connect directly to broader course themes like statistical reasoning, data interpretation, and quantitative literacy.

Think of graph types as tools in a toolkit—each one is designed for a specific job. A hammer isn't better than a screwdriver; it just solves a different problem. Your goal isn't to memorize definitions but to understand what question each graph type answers. When you see a dataset, ask yourself: Am I showing change over time? Comparing categories? Revealing a relationship between variables? The answer tells you which graph to reach for.

---

Graphs for Showing Change Over Time

These graphs excel at revealing trends, patterns, and cycles in data collected at regular intervals. The key mechanism: the horizontal axis represents time, allowing viewers to track how values evolve sequentially.

Line Graph

• Connects data points with straight lines—the slope between points instantly communicates whether values are rising, falling, or stable
• Best for continuous data over time, such as stock prices, temperature readings, or population growth
• Reveals patterns like seasonal cycles, long-term trends, and sudden changes that might be missed in a table

Area Graph

• Fills the space below the line—this visual weight emphasizes the magnitude of values, not just their direction
• Effective for cumulative totals or showing how multiple data series stack up over time
• Highlights volume and scale, making it easier to see the "size" of change rather than just the trend

---

Graphs for Comparing Categories

When you need to answer "which category is biggest?" or "how do groups differ?", these graphs organize data by discrete categories rather than continuous variables. The visual comparison happens through length, height, or position.

Bar Graph

• Uses rectangular bars where length corresponds to value—our eyes naturally compare bar heights quickly and accurately
• Works for any categorical data, from survey responses to sales by region to test scores by class
• Can be vertical or horizontal—horizontal bars work better when category names are long

Stacked Bar Graph

• Layers multiple data series within each bar—shows both the total and the composition of each category
• Reveals part-to-whole relationships while still allowing comparison across groups
• Trade-off to remember: easy to compare totals and bottom segments, harder to compare middle segments accurately

---

Graphs for Part-to-Whole Relationships

These visualizations answer the question: "What fraction or percentage does each category represent?" The underlying principle: the whole is divided into proportional pieces that must sum to 100%.

Pie Chart

• Divides a circle into slices—each slice's angle and area represent that category's proportion of the total
• Most effective with 2-6 categories; more slices make comparisons difficult and the chart cluttered
• Limitation: hard to compare similar-sized slices precisely, and impossible to show change over time

---

Graphs for Showing Relationships Between Variables

These graphs plot data points to reveal correlations, clusters, and patterns between two or more variables. The core mechanism: position on a Cartesian plane encodes the values of each variable simultaneously.

Scatter Plot

• Plots individual data points on an $x$-$y$ plane—each point represents one observation with two measured values
• Reveals correlation strength and direction: positive (upward trend), negative (downward trend), or none (random scatter)
• Exposes outliers and clusters that might indicate subgroups or data errors

Bubble Chart

• Adds a third variable through bubble size—transforms a 2D scatter plot into a visualization of three dimensions
• Useful for complex datasets where you need to show relationships among multiple variables simultaneously
• Interpretation challenge: our eyes struggle to compare circle areas accurately, so use for general patterns, not precise readings

---

Graphs for Showing Data Distribution

Distribution graphs answer: "How is my data spread out? Where do values cluster? Are there outliers?" These visualizations focus on frequency—how often different values or ranges occur.

Histogram

• Groups continuous data into intervals (bins)—bar height shows how many data points fall within each range
• Reveals distribution shape: normal (bell curve), skewed left/right, uniform, or bimodal
• Critical distinction from bar graphs: histograms show continuous data with no gaps between bars; bar graphs show categorical data with gaps

Box Plot (Box-and-Whisker Plot)

• Summarizes data with five numbers: minimum, first quartile ($Q_1$), median, third quartile ($Q_3$), and maximum
• Shows spread and skewness at a glance—the box contains the middle 50% of data (the interquartile range)
• Flags potential outliers as individual points beyond the whiskers, making comparison between groups efficient

Stem-and-Leaf Plot

• Preserves original data values while showing distribution—the "stem" is the leading digit(s), the "leaf" is the final digit
• Best for small datasets (roughly 15-50 values) where you want to see both shape and actual numbers
• Quick to construct by hand, making it useful for exploratory analysis before choosing a more polished visualization

---

Quick Reference Table

---

Self-Check Questions

1. You have monthly sales data for three products over two years. Which graph type would best show both the individual trends and how they compare to each other? Why might a line graph work better than an area graph here?

2. What do histograms and bar graphs have in common visually, and what critical difference determines when to use each one?

3. A researcher wants to show the relationship between study hours, test scores, and class size for 30 students. Which graph type accommodates all three variables, and what limitation should they keep in mind?

4. Compare box plots and histograms: if you needed to compare the test score distributions of five different classes on a single page, which would you choose and why?

5. A pie chart shows budget allocation across eight departments, but two slices look nearly identical. What alternative visualization would make the comparison clearer, and what information would it add that the pie chart cannot show?