Graphing Data by Hand: Scatter Plot Essentials

Example experiment: vary the amount of fertilizer (independent variable) and measure plant growth (dependent variable).
Independent variable (x): Fertilizer amount (in grams).
Dependent variable (y): Plant growth (in centimeters).
Purpose: use a scatter plot to examine the correlation between two numeric data sets.

Choose a scatter plot because you’re looking at the relationship between two numerical values (x and y).
In a scatter plot, each point represents a single observation (one plant under one fertilizer amount).

X-axis: Fertilizer amount (grams).
Y-axis: Plant growth (centimeters).
Rationale: the thing you change goes on the x-axis; the thing you measure goes on the y-axis.
Axis labels should explicitly state the variable and units, e.g.,
- Fertilizer (grams)
- Plant growth (centimeters)

Use a descriptive title that includes context when possible (e.g., the plant type or time period) rather than a vague title like "Plant growth".
The title should convey the relationship being examined: e.g., "Relation of Fertilizer to Plant Growth (Plant type X, 2 weeks)".
Labels should be legible and informative so you can infer the axes without referring back to the data source.
Use all of the graph paper space to maximize readability and pattern detection.

Dataset details from the example:
- X-values (fertilizer): range from 2 to 12 grams.
- Y-values (growth): range from 18 to 98 cm.
Total number of squares counted on the x-axis: 21 squares.
Total number of squares counted on the y-axis: 15 squares.
Calculating the x-scale (horizontal, for fertilizer):
- Range on x: $ext{range}<em>x = x</em>{ ext{max}} - x_{ ext{min}} = 12 - 2 = 10$
- If you had Nx = 21 squares, then the ideal step per square would be $$ riangle x = rac{ ext{range}x}{N_x} = rac{10}{21} \