Advanced Charts - Part 1

Scatter plots require two numerical variables—one for the X-axis and one for the Y-axis. Each point plots where these two values intersect. For example, height (X) and weight (Y) are both numerical, allowing us to plot each person as a point. Categorical variables like "product type" don't work on scatter plots because they can't be positioned on a continuous numerical scale, as explained in Chapter 16.

An upward slope from left to right indicates a positive correlation—as one variable increases, the other tends to increase as well. In this case, more study hours are associated with higher test scores. However, remember that correlation does not prove causation; while the pattern is clear, other factors might also influence test scores. Chapter 16 emphasizes this critical distinction between correlation and causation.

The key difference is that area charts fill the space between the line and the baseline (usually zero), creating a shaded area that emphasizes the magnitude or volume of the data. Line charts just show the line itself, focusing attention on the trend. Both show the same trend, but the filled area makes the total quantity more visually prominent, as discussed in Chapter 16's comparison of these two chart types.

This is a classic example of spurious correlation—a third variable (hot weather/temperature) drives both ice cream sales and swimming activity, which increases drowning incidents. Correlation does NOT equal causation. Just because two variables move together doesn't mean one causes the other. Chapter 16 warns about this common misinterpretation and emphasizes the importance of looking for confounding variables.

A 100% stacked area chart normalizes all data to show proportions that always add up to 100% at each point in time. This means you lose information about absolute totals but gain the ability to easily compare how the composition changes over time. For example, you can see that Product A's market share grew from 30% to 45%, even though you don't see the actual dollar amounts. Chapter 16 explains when this trade-off is worthwhile.

Add a trend line when there's a clear linear relationship that the line can help illustrate. However, don't force a trend line onto data with no correlation or non-linear patterns—it can be misleading. The trend line should enhance understanding, not create a false impression of a relationship that doesn't exist. As covered in Chapter 16, trend lines are powerful when used appropriately but can deceive when applied indiscriminately.

A stacked area chart perfectly shows both the trend over time (quarters) and the composition (how each product contributes). The total height shows combined revenue, while the colored layers show each product's contribution. This is exactly the use case for stacked area charts, as explained in Chapter 16—showing how parts contribute to a whole while also tracking changes over time.

When values go negative, the filled area extends below the baseline, creating visual confusion about what the shading represents. It becomes unclear whether you're looking at magnitude above or below zero. Use line charts instead for data that crosses the zero line, as they handle positive and negative values clearly without the ambiguity of filled areas. This best practice is highlighted in Chapter 16's discussion of when NOT to use area charts.

Each point on a scatter plot represents one observation (like one student, one product, or one day) with values for both the X and Y variables. For example, in a scatter plot showing study hours vs. test scores, each dot represents one student's study time and their corresponding test score. This individual-level representation allows you to see the overall relationship while also identifying outliers or special cases, as explained in Chapter 16.

Area charts are best when you want to emphasize the magnitude or volume of change over time, not just the trend. The filled area makes the total amount more visually prominent than a line alone. They're especially useful for cumulative data like total users over time or total sales to date. However, as covered in Chapter 16, avoid area charts for data with negative values since the filled area below the baseline becomes confusing.