Visualizing and Describing Data
Outline
- use appropriate graphs to display and describe quantitative and categorical data
- describe the distribution of a variable
- make graphs that describe the relationship between 2 or more variables
Data Visualization displaying data in ways that make patterns more noticeable
Data Visualization Guide Table
| Graph Type | Variables | Purpose | Pros | Cons |
|---|---|---|---|---|
| Bar Graph | 1 categorical | Display frequency/counts of categories | Simple, intuitive for comparing groups | Limited to counts; hides distribution details |
| Histogram | 1 quantitative | Show distribution shape, center, spread, outliers | Reveals patterns (skew, modes) | Bin size choice affects interpretation |
| Boxplot | 1 quantitative | Summarize center (median), spread (IQR), and outliers | Robust to outliers; compact visualization | Hides multimodality; less detail than histograms |
| Stacked/Dodged Bar | 2 categorical | Compare proportions/counts across categories | Shows subgroup breakdowns | Stacked: Hard to compare subgroups; Dodged: Space-intensive |
| Scatterplot | 2 quantitative | Explore relationships (form, direction, strength) | Reveals trends, clusters, outliers | Overplotting with large datasets |
| Side-by-Side Boxplots | 1 quantitative + 1 categorical | Compare distributions across categories | Highlights differences in medians/IQR | Simplifies distribution details |
| Faceted/Colored Plot | 3+ variables (mix of types) | Multivariate analysis (e.g., relationships across subgroups) | Incorporates multiple variables in one plot | Complexity; may become cluttered |
Association
When there seems to be some connection between two variables (knowing about one variable tells us about the other), we say they are associated.
If there does not seem to be a relationship between the variables, we say they are independent.