Scatter Plot Real-Life Examples — 8 Use Cases Across Business, Science, and Sports

The fastest way to understand scatter plots is to see them applied to situations you care about. Textbook examples like "shoe size vs. IQ" are abstract. Real-world examples — marketing spend vs. signups, training hours vs. race times, advertising dollars vs. ticket sales — stick because you have intuitions about whether the correlation should exist.

Below are 8 scatter plot examples you can recreate in the free scatter plot maker. Copy the sample data, paste it in, and see the chart. Notice how the pattern matches (or contradicts) your intuition.

Example 1: Ad Spend vs. Conversions (Strong Positive)

Scenario: A small business tracked monthly ad spend and resulting conversions over 12 months.

500, 12
750, 18
1000, 22
1200, 25
1500, 31
1800, 38
2000, 41
2200, 43
2500, 48
2800, 52
3000, 55
3200, 58

Expected pattern: Strong positive correlation. More spend, more conversions.

What you learn: When the R-squared is above 0.95, the relationship is very predictable. You can use the regression equation to forecast what spending $4,000 might yield (though remember: extrapolation beyond your data range gets unreliable).

Example 2: Study Hours vs. Exam Score (Moderate Positive)

Scenario: 15 students reported hours studied and their final exam score.

2, 58
3, 62
4, 71
5, 68
6, 75
7, 80
8, 77
9, 84
10, 88
12, 85
15, 91
18, 94
20, 89
22, 95
25, 92

Expected pattern: Moderate positive. More study generally means higher scores, but other factors (talent, prior knowledge, test anxiety) add noise.

What you learn: Even "obvious" relationships have scatter. A 15-hour studier scored 91, but a 20-hour studier scored 89. Studying is a factor, not the only factor. This is how most real-world relationships look — strong enough to matter, not perfect enough to predict individual cases.

Example 3: Weekly Training Miles vs. 5K Time (Strong Negative)

Scenario: 12 runners reported weekly mileage and their most recent 5K time (in minutes).

5, 28.5
8, 26.2
10, 25.8
12, 24.3
15, 23.1
18, 22.4
20, 21.8
25, 21.0
30, 20.3
35, 19.8
40, 19.2
45, 18.9

Expected pattern: Strong negative. More miles, faster times.

What you learn: Negative correlation does not mean weak correlation. R-squared above 0.9 with a negative slope is just as strong as with a positive slope. The negative sign tells direction, not strength.

Example 4: House Square Footage vs. Price (Strong Positive)

Scenario: 10 recent home sales in a neighborhood.

1200, 285000
1450, 310000
1600, 340000
1800, 375000
2000, 395000
2200, 420000
2400, 445000
2600, 485000
2800, 510000
3200, 575000

Expected pattern: Strong positive. Bigger houses cost more.

What you learn: The slope (from the equation y = mx + b) tells you approximately what each additional square foot adds to the price. For this neighborhood, the slope suggests each extra sqft adds ~$145 to the price.

Sell Custom Apparel — We Handle Printing & Free Shipping

Example 5: Shoe Size vs. Annual Income (No Correlation)

Scenario: Survey data from 15 adults.

7, 52000
8, 78000
9, 45000
10, 92000
10.5, 65000
11, 58000
8.5, 71000
9.5, 83000
12, 49000
7.5, 95000
11.5, 67000
9, 88000
10, 54000
8, 102000
11, 73000

Expected pattern: None. Shoe size has no relationship to income.

What you learn: R-squared close to 0 confirms no linear relationship. The trend line the tool draws through this data is technically real but meaningless. In analysis writeups, note low R-squared and avoid implying the line says something it does not.

Example 6: Daily Temperature vs. Ice Cream Sales (Strong Positive)

Scenario: An ice cream shop tracked daily high temperature (Fahrenheit) and sales for 14 days in summer.

72, 185
75, 210
78, 245
81, 278
83, 295
85, 320
87, 345
89, 372
91, 395
93, 418
95, 432
92, 405
88, 360
84, 305

Expected pattern: Strong positive. Hotter days, more ice cream sales.

What you learn: This is a classic example used to teach correlation vs. causation. Temperature correlates with sales, and in this case causation is plausible (hot people want ice cream). But correlation alone does not prove causation — always consider whether a third variable could be driving both.

Example 7: Portfolio Risk vs. Annual Return (Moderate Positive)

Scenario: 12 investment portfolios with their risk score (standard deviation) and 5-year average annual return.

5, 4.2
7, 5.8
8, 5.5
10, 7.1
12, 8.4
14, 9.2
15, 8.7
18, 11.3
20, 10.8
22, 12.5
25, 11.9
28, 13.7

Expected pattern: Moderate positive. Higher risk generally earns higher returns, but the relationship is noisy — not every risky portfolio outperforms.

What you learn: The risk-return tradeoff is real but imperfect. Some risky portfolios underperform safer ones. The scatter shows why diversification matters — you cannot just pick the highest-risk option and expect the highest return.

Example 8: Plant Height vs. Daily Sunlight Hours (Very Strong Positive)

Scenario: 10 tomato plants grown with different daily sunlight exposure, measured after 30 days.

2, 4.5
3, 8.2
4, 12.1
5, 15.8
6, 20.3
7, 24.9
8, 28.4
9, 31.8
10, 34.5
12, 38.2

Expected pattern: Very strong positive. Controlled experiment with a clear biological mechanism.

What you learn: Experiments with controlled variables produce the cleanest scatter plots — R-squared above 0.99 is common. Real-world observational data is always messier. Compare this example to the study hours vs. exam score example (moderate correlation) — same positive direction, very different strength, because one is controlled and one is observational.

Try these examples in the scatter plot maker. The more patterns you see, the faster you can read a new scatter plot at a glance. For more on interpreting what you see, check our correlation types guide.

Frequently Asked Questions

What is a good R-squared value for real-world data?

For controlled experiments, R-squared above 0.9 is common. For real-world observational data, 0.6-0.8 is often strong. Below 0.3 usually means the relationship is weak or non-linear. Context matters — a 0.4 R-squared in social science might be considered strong, while a 0.4 in physics would be poor.

Does correlation mean one variable causes the other?

No. Correlation shows two variables move together. Causation requires additional evidence: a plausible mechanism, experimental control, time ordering (cause precedes effect), and ruling out third variables. Ice cream sales correlate with drowning deaths (both rise in summer), but one does not cause the other — temperature is the hidden common cause.

How many data points do I need for a meaningful scatter plot?

The general rule is at least 15-20 points for a reliable visual pattern. Fewer than 10 points can be misleading — a few coincidental values can drive R-squared high or low. For statistical significance testing, even more points are recommended.