01 / 21

Semi-log
Plots

Learn how logarithmic axes transform exponential data into straight lines, making patterns easy to identify and analyze.

AP Precalculus

What Is a Semi-log Plot?

A semi-log plot (also called a semi-logarithmic plot) is a graph where one axis uses a logarithmic scale and the other uses a standard linear scale. In AP Precalculus, the most common version places \(\log y\) on the vertical axis and \(x\) on the horizontal axis.

Why Use a Logarithmic Axis?

Exponential data spans huge ranges
Values like 10, 100, 10,000, and 1,000,000 all need to fit on one graph.
Curves become lines
A logarithmic axis "straightens out" exponential growth or decay, revealing the trend.
Easy model identification
If the semi-log plot is linear, the data is exponential. If it curves, the data is not.

Standard vs. Semi-log View

Standard Plot
\(x\)
\(y\)
0 200 400 600 800 1000 1 2 3 4 Curved! Semi-log Plot
\(x\)
\(\log y\)
0 0.5 1.0 1.5 2.0 2.5 3.0 1 2 3 4 Straight!

The Core Algebraic Idea

Start with the exponential model and take the logarithm of both sides:

If we let \(Y = \log y\), this becomes:

Reading Slope and Intercept

Slope \(m\)
Equals \(\log b\). To find the base: \(b = 10^{m}\). A steeper line means faster exponential growth.
y-intercept \(c\)
Equals \(\log a\). To find the initial value: \(a = 10^{c}\). This is the value of \(y\) when \(x = 0\).

A cell culture where more cells create more growth events

A researcher looks through a microscope beside a closed culture vessel,sample containers and an illustrative cell inset.The line log₁₀P = 0.3t + 2 gives P = 100×10^(0.3t). Its hourly growth factor is about 1.995.

Semi-log source preset

The line log₁₀P = 0.3t + 2 gives P = 100×10^(0.3t). Its hourly growth factor is about 1.995.The line log₁₀P = 0.3t + 2 gives P = 100×10^(0.3t). Its hourly growth factor is about 1.995.

Semi-log vs. Log-log Plots

Semi-log Log-log
Axes \(\log y\) vs. \(x\) \(\log y\) vs. \(\log x\)
Linearizes Exponential models \(y = ab^{x}\) Power models \(y = ax^{n}\)
Slope gives \(\log b\) (base of exponent) \(n\) (power/exponent)
Intercept gives \(\log a\) (initial value) \(\log a\) (coefficient)

Checking Model Fit with Residuals

After fitting a line to the semi-log data, compute residuals: the differences between observed \(\log y\) values and the values predicted by the line.

1
Random residuals: scattered above and below zero with no pattern: the exponential model is a good fit.
2
Patterned residuals: a curve or systematic trend in the residual plot: the data may not be truly exponential.

How to Build a Semi-log Model

1
Compute \(\log y\) for every data point (keep \(x\) unchanged).
2
Plot the points \((x,\,\log y)\) on a standard set of axes.
3
Fit a line: find slope \(m\) and intercept \(c\) using two points or a calculator.
4
Recover the model: \(a = 10^{c}\) and \(b = 10^{m}\), giving \(y = ab^{x}\).
5
Check residuals: plot \(\log y_{\text{observed}} - \log y_{\text{predicted}}\) and look for random scatter.

Example 1: Bacteria Growth

A culture starts with 50 bacteria and doubles every hour. Data after 0, 1, 2, 3, 4, and 5 hours:

\(x\) (hours) \(y\) (count) \(\log y\)
0501.699
11002.000
22002.301
34002.602
48002.903
516003.204

Example 1: Finding the Model

\(x\) (hours)
\(\log y\)
1.5 2.0 2.5 3.0 3.5 0 1 2 3 4 5 slope = 0.301 intercept = 1.699

Example 2: Radioactive Decay

A substance has a half-life of 3 days. Starting mass: 800 g. A semi-log plot of \(\log y\) vs. \(x\) (days) passes through the points \((0,\,2.903)\) and \((3,\,2.602)\).

1
Slope: \(m = \frac{2.602 - 2.903}{3 - 0} = \frac{-0.301}{3} \approx -0.1003\)
2
Base: \(b = 10^{-0.1003} \approx 0.794\)
3
Initial value: \(a = 10^{2.903} \approx 800\)
4
Model: \(y = 800 \cdot (0.794)^{x}\), which halves every 3 days.

Example 3: Natural Log Version

Some semi-log plots use \(\ln y\) instead of \(\log_{10} y\). The algebra is the same but with base \(e\):

Suppose a semi-log plot (\(\ln y\) vs. \(x\)) has slope \(k = 0.05\) and intercept \(\ln a = 6.91\). Then:

When the Semi-log Plot Curves

\(x\)
\(\log y\)
Exponential data Non-exponential data

Common Mistakes to Avoid

✕ Forgetting to take the log
Plotting raw \(y\) values on the vertical axis instead of \(\log y\) defeats the entire purpose of a semi-log plot.
✕ Mixing log bases
Using \(\log_{10}\) for the slope but \(e\) for the model. If slope = \(\log_{10} b\), recover \(b = 10^{m}\), not \(e^{m}\).
✕ Applying the wrong plot type
Semi-log plots linearize exponential data (\(y=ab^x\)). Log-log plots linearize power data (\(y=ax^n\)). Using the wrong one will lead to incorrect conclusions.

Summary at a Glance

Feature Semi-log Plot
Axes \(\log y\) (vertical) vs. \(x\) (horizontal)
Data model \(y = ab^{x}\) (exponential)
Linear form \(\log y = (\log b)x + \log a\)
Slope means \(\log b\): multiplicative rate per unit \(x\)
Intercept means \(\log a\): initial value (at \(x = 0\))
Positive slope Exponential growth (\(b > 1\))
Negative slope Exponential decay (\(b < 1\))
Curved data Not exponential: try another model

Question 1 of 5

On a semi-log plot, \(\log y\) is plotted against \(x\) and the data forms a straight line. What type of function does the original data follow?

Question 2 of 5

A semi-log regression line has slope \(m = 0.477\) and y-intercept \(c = 2\). What is the exponential model?

Question 3 of 5

On a semi-log plot, the regression line has a negative slope. What does this tell you about the original data?

Question 4 of 5

Data points on a semi-log plot form a curve rather than a line. Which conclusion is most appropriate?

Question 5 of 5

A semi-log plot using \(\ln y\) vs. \(x\) has slope \(k = 0.03\) and intercept \(\ln a = 4.605\). What is the model?