Data Visualization: Tufte's lie factor

Edward Tufte's lie factor compares how much a chart's graphic changes with how much the data change. Compute it and judge the chart.

What to build

A program that computes the lie factor of a chart.

Requirements

  • Print data change X% and graphic change Y%, the relative changes as percentages to one decimal.
  • Then print lie factor Z (graphic change ÷ data change, two decimals). If the data change is 0, the lie factor is 1.00 when the graphic change is also 0, otherwise print lie factor infinite.
  • Finally print honest if the lie factor is between 0.95 and 1.05 inclusive, otherwise misleading.

Input

One line: the two data values, then the two drawn sizes.

Output

Four lines.

Constraints

All numbers are over 0 and at most 10⁶.

Example

Input:

18 27.5 0.6 5.3

Output:

data change 52.8%
graphic change 783.3%
lie factor 14.84
misleading

Why it works

Comparing relative changes removes the units of data and graphic, so a ratio of 1 means the picture changes exactly as much as the data, and anything else means it overstates or understates the change.

Examples

Example 1 · Input
18 27.5 0.6 5.3
Expected output
data change 52.8%
graphic change 783.3%
lie factor 14.84
misleading
Example 2 · Input
10 20 1 2
Expected output
data change 100.0%
graphic change 100.0%
lie factor 1.00
honest

Prompts inside input() are ignored when checking, so input("n: ") is fine.

Hints

Hint 1

Each effect is |second − first| ÷ first.

Hint 2

Handle a data effect of 0 separately to avoid dividing by zero.

Hint 3

A lie factor from 0.95 to 1.05 counts as honest.

Solution

Show one correct solution (try it yourself first)
d1, d2, g1, g2 = map(float, input().split())
data = abs(d2 - d1) / d1
graphic = abs(g2 - g1) / g1
print(f"data change {data * 100:.1f}%")
print(f"graphic change {graphic * 100:.1f}%")
if data == 0:
    lf = 1.0 if graphic == 0 else None
else:
    lf = graphic / data
if lf is None:
    print("lie factor infinite")
    print("misleading")
else:
    print(f"lie factor {lf:.2f}")
    print("honest" if 0.95 <= lf <= 1.05 else "misleading")

Lesson: Misleading charts →