Semi-log Plots and Linearization

When scientists measure quantities that grow or decay exponentially — like bacterial populations, radioactive decay, or earthquake intensities — plotting raw data on ordinary axes produces a wildly curved graph that is difficult to analyze. Semi-log plots solve this problem by rescaling one axis using a logarithm, transforming an exponential curve into a straight line that is easy to interpret and work with. The key insight is algebraic: if y = ab^x, then taking the logarithm of both sides gives log(y) = log(a) + x·log(b), which is a linear equation in x and log(y). This process — rewriting a nonlinear relationship as a linear one — is called linearization. On a semi-log plot the vertical axis shows log(y) while the horizontal axis shows x normally, so exponential data appears as a perfectly straight line whose slope encodes the growth rate. Semi-log plots are used throughout science and engineering. Geologists use them to date rock samples, epidemiologists use them to track disease spread, and engineers use them to test electronic components. Mastering this technique connects logarithmic algebra directly to real-world data analysis, giving you a powerful tool for identifying and quantifying exponential relationships.

Key takeaways
  • Start with the Exponential Model
  • Apply a Logarithm to Both Sides
  • Recognize the Linear Form
  • Read the Slope and Intercept
  • Interpreting the Slope Physically
  • Checking Whether Data Is Truly Exponential

Start with the Exponential Model

Suppose data follows y = a · b^x, where a is the initial value (y-intercept) and b is the base controlling growth or decay. On a standard xy-plot this produces a curved graph — hard to fit by eye and difficult to use for predictions.

Apply a Logarithm to Both Sides

Take log₁₀ of both sides: log(y) = log(a · b^x). Using logarithm rules, log(ab^x) = log(a) + log(b^x) = log(a) + x·log(b). This gives log(y) = log(b)·x + log(a).

Recognize the Linear Form

Let Y = log(y). Then Y = log(b)·x + log(a) is exactly Y = mx + c with slope m = log(b) and y-intercept c = log(a). Plotting Y versus x gives a straight line. This is the semi-log linearization.

Read the Slope and Intercept

From the straight line on the semi-log plot, identify the slope m = log(b) and the Y-intercept c = log(a). Then recover the original parameters: b = 10^m and a = 10^c. These give you the explicit exponential formula describing your data.

Interpreting the Slope Physically

Because m = log(b), the slope tells you the factor by which y multiplies for each unit increase in x. If m = 0.3, then b = 10^0.3 ≈ 2.0, meaning y approximately doubles for every unit increase in x. A negative slope signals exponential decay. The steeper the slope, the faster the growth or decay.

Checking Whether Data Is Truly Exponential

Semi-log plots also serve as a diagnostic. If real data plotted with log(y) on the vertical axis falls approximately along a straight line, that is strong evidence the underlying relationship is exponential. If the points curve on the semi-log plot, the relationship is not purely exponential and a different model may be needed.

The Key Equation

log⁡(y)=log⁡(b)⋅x+log⁡(a)\log(y) = \log(b) \cdot x + \log(a)
  • y — Dependent variable (varies): The quantity being measured; must be positive for the logarithm to be defined.
  • x — Independent variable (varies): The input variable plotted on the horizontal axis with a linear (uniform) scale.
  • a — Initial value (same as y): The value of y when x = 0; equals 10^(y-intercept of the semi-log plot).
  • b — Base / growth factor (dimensionless): The multiplicative factor per unit increase in x; equals 10^(slope of semi-log line).
  • \log(b) — Slope of semi-log line (dimensionless): Positive for exponential growth, negative for exponential decay.

Worked Example

Problem. A biologist records the following bacteria counts N (thousands) at times t (hours): t = 0, N = 2; t = 1, N = 6; t = 2, N = 18; t = 3, N = 54. Plot log(N) vs t, find the slope and intercept, and write the exponential model.

N(t)=2⋅3tN(t) = 2 \cdot 3^{t}

Answer. N(t) = 2 · 3^t thousand bacteria; the population triples each hour.

Figure
Diagram
Semi-Log Plot: log₁₀(N) vs tTime t (hours)log₁₀(N)0.30.81.31.70123(0, 0.301)(1, 0.778)(2, 1.255)(3, 1.732)slope = 0.477Δt=1Δlog N=0.477
Origins · 1 minJohn Napier and how this was discovered
JN
John Napier
1550–1617 · Scottish

John Napier invented logarithms around 1614, publishing his tables in Mirifici Logarithmorum Canonis Descriptio. His invention transformed astronomy, navigation, and engineering by converting tedious multiplication and division into addition and subtraction — exactly the property that makes semi-log linearization possible today. Napier spent 20 years computing his logarithm tables by hand. He also invented 'Napier's bones', a set of carved rods used as a mechanical calculator, and was rumored by neighbors to practice witchcraft because of his unusual scientific experiments.

  • Properties of Logarithms — The linearization step uses log(ab^x) = log(a) + x·log(b), which requires fluency with product and power rules for logarithms.
  • Exponential Functions — Semi-log plots are specifically designed to linearize exponential functions y = ab^x; understanding their behavior is essential context.
  • Linear Regression — After linearizing data via a semi-log plot, linear regression is used to find the best-fit line through the transformed points.
  • Logarithmic Scales (Richter, Decibels, pH) — Semi-log plots use the same rescaling idea as the Richter scale and decibel system — a logarithmic axis compresses wide-ranging data into a readable form.