Graphs

Graphs: An essential tool for data representation; they offer the advantage of being interpretable but can also misrepresent data.
- Often the most interpretable form but not necessarily the most accurate.
- Can manipulate, distort, and misrepresent data through various means.

Discrete Data and Variables:
- Nominal or Categorical Level Data:
  - Finite categories/distinct groups.
  - Can count responses; must be whole numbers.
  - Examples:
  - US census regions
  - Biological sex
  - Class standing
  - Marital status
  - Payment methods
  - Movie genres
  - Streaming services
  - Grades.
Continuous Data and Variables:
- Ordinal, Interval, Ratio Level Data:
  - Infinite number of values between any two values; can be any numerical value.
  - Examples:
  - Age
  - Income in dollars
  - Height
  - Weight
  - Grades in a class.

For Discrete Data (Categorical Variables):
- Pie Charts:
  - Display the percentage of cases in each category.
- Bar Charts:
  - Present percentage or number of cases; it’s possible to show multiple categories simultaneously.
- Pictograms:
  - Use images to represent data, often used in magazines.
For Continuous Data:
- Histograms (Frequency Curves):
  - Present continuous data and display data distributions.
  - X-Axis: Values (e.g., defects per hour).
  - Y-Axis: Number of cases within intervals.
- Scatter Plots:
  - Show relationships between two continuous variables; each data point is represented by a dot.
- Line Graphs:
  - Display continuous variable values over time, suitable for showing trends.

Efficient: Balance simplicity with the necessary information displayed.
Accurate: Data must be represented without distortion.
Aesthetically Pleasing: Good use of color, font, and size; the graph should be “user-friendly.”
Accessible: Information should be easy to understand without excessive cognitive effort.

Pie Charts Characteristics:
- Must sum to 100%.
- Typically represent discrete categories such as fruits or political candidates.
Bar Charts Characteristics:
- Show percentage or number of cases in categories; height indicates percentage.
- Can be oriented vertically or horizontally.
Pictograms Characteristics:
- Similar to bar charts but utilize images instead of bars for representation.

Histograms:
- Similar to bar charts; display distributions of continuous data.
- Example:
  - X-Axis: Defects per hour, Y-Axis: Number of cases.
Cumulative Frequency Curves:
- Demonstrate frequency distribution; helpful for showing cumulative totals over intervals.
Scatter Plots:
- Each data point represented as a dot, illustrating relationships between two continuous variables.
Line Graphs:
- Useful for visualizing trends over time with continuous data.

Omitting Labels:
- No labels on axes can mislead viewers about the data’s context.
- Example: A chart displaying membership growth without a Y-axis label.
Y-Axis Manipulation:
- Not starting the Y-axis at zero leads to distortion of data representation; can exaggerate perceived changes.
- Using small units or non-equal intervals can artificially inflate or minimize apparent changes.
Inaccurate Measurements:
- Height of bars not consistently representing values on the Y-axis; misleading visual encoding of data.
- Improper scale or units on the X-axis can misinform the audience (e.g., inflated growth rates).
Cumulative Curves:
- Present data in a way that suggests misleading accumulation; often provides an inflated sense of growth or presence of a metric.
- Example of misleading cumulative iPhone sales data: could show growth inaccurately over a period, especially if axes not labeled or improperly scaled.

Graphs: Although they simplify data understanding, they can be misleading and manipulated.
Types of Graphs: Differences between discrete vs. continuous data must be recognized when selecting graph types.
Techniques for Manipulation: Important to recognize how graphical misrepresentation can occur through unlabeled axes, nonsensical scaling, and cumulative representations.