Free graph maker & data converters

Log Scales: Semi-Log vs Log-Log Graphs

By Plotiq Editorial Team · Published 14 November 2026 · Updated 14 November 2026 · 4 min read

A logarithmic axis turns multiplicative change into equal steps, so data that grows or decays by a constant factor becomes a straight line. This guide explains how to read a log axis, when to use semi-log or log-log, and how to pull a growth rate or exponent out of the slope.

What a logarithmic axis does

On an ordinary linear axis, equal distances mean equal differences: the step from 10 to 20 is the same length as the step from 90 to 100. On a logarithmic axis, equal distances mean equal ratios: the step from 1 to 10 is the same length as the step from 10 to 100 and from 100 to 1,000. Each such step is called a decade.

There is no zero on a log axis, because the logarithm of zero is undefined, and negative values cannot be plotted either. Tick labels typically show powers of ten (1, 10, 100), with minor ticks at 2, 3, 4 and so on between them, getting closer together as they approach the next decade. If the minor ticks are not evenly spaced, that is the log scale working correctly.

Semi-log: one axis logarithmic

A semi-log plot makes only one axis logarithmic, usually Y. It is the right choice when Y changes by a constant factor per unit of X, which is exponential growth or decay. If y = a × e^(kx), then ln y = ln a + kx, which is a straight line in X. So a straight line on a semi-log plot signals exponential behaviour, and its steepness is the rate.

Typical examples are bacterial growth, radioactive decay, compound interest, drug concentration in blood after a dose, and early epidemic spread. In each case the quantity changes by the same percentage in each time step, and the curve on a linear axis is hard to judge by eye.

Worked example: bacterial growth

The optical density at 600 nm (OD₆₀₀) of a bacterial culture rises over seven hours: 0.020, 0.034, 0.061, 0.104, 0.180, 0.312, 0.540, 0.931. On a linear axis it looks like a steepening curve. Each reading is about 1.7 times the previous one, a constant ratio, so on a log Y axis the points fall on a straight line, as in the figure.

The growth rate comes from the slope of ln(OD) against time. Between 0 h and 7 h the OD rises by a factor of 0.931 / 0.020 = 46.6, and ln(46.6) = 3.84, so the rate constant is 3.84 / 7 = 0.55 h⁻¹. The doubling time is ln 2 / 0.55 = 1.3 hours. That is a number you can compare between strains or growth media, which is the whole point of the plot.

012345670.010.11Time (h)OD600
Exponential growth is a straight line when the Y axis is logarithmic. Made with the Log Graph Maker.

Log-log: both axes logarithmic

A log-log plot makes both axes logarithmic. It reveals power-law relationships, y = a × x^b, because taking logs gives ln y = ln a + b ln x, a straight line whose slope is the exponent b. Examples include how metabolic rate scales with body mass, how the period of a pendulum or orbit scales with length, drag force against speed, and the frequency response of filters in engineering.

Read the exponent from the slope as long as both axes use the same decade length, or calculate it from two points: b = log(y₂ / y₁) / log(x₂ / x₁). A slope of 1 means y is proportional to x, 2 means it goes with the square, and 0.75 is the well-known quarter-power scaling of animal metabolism.

Linear, semi-log or log-log: how to decide

Try each one and see which gives a straight line, because straight lines are easy to judge and fit:

Straight line on…RelationshipSlope tells you
Linear X, linear Yy = mx + bChange in y per unit x
Linear X, log Yy = a·ekxRate constant k (growth or decay)
Log X, linear Yy = a + b·log xChange in y per decade of x
Log X, log Yy = a·xbThe exponent b

Pitfalls to avoid

Zero and negative values cannot be shown; if your data include zeros, a log axis is the wrong tool or the zeros need a documented treatment. Log axes compress large values, so they can make a big absolute change look small; always label the axis as logarithmic and keep the tick labels. Do not fit a straight line by eye on a log plot and quote the slope as if it were a linear one. Fit the model properly, using the exponential or power trendline, which calculates the parameters for you. And remember that an R² computed on logged data weights relative errors, not absolute ones, so it is not directly comparable with an R² from a linear fit.

Frequently asked questions

Why can’t I plot zero on a log scale?
The logarithm of zero is undefined: it tends to minus infinity. Values must be positive. Either remove the zeros, add a small documented offset, or use a linear axis.
What is the difference between ln and log10?
They differ by a constant factor (ln x = 2.303 × log₁₀ x). The straight-line shape is identical; only the numerical value of the slope changes. Use the one that matches your formula, and state it.
How do I know if data are exponential or a power law?
Plot both. A straight line on a semi-log plot indicates exponential; a straight line on a log-log plot indicates a power law. Data spanning less than a decade or two cannot reliably tell them apart.
Do I need to log-transform my data first?
No. Switching the axis to log scale does it for you and labels the ticks in the original units. Transform the data only when you need the logged values for a calculation.

Keep reading

  • Line of Best Fit and R²: What They Really Tell You – How least-squares finds a line of best fit, what R² does and does not tell you, how to read residuals and when to use a curve.
  • Which Chart Should I Use? – A practical decision guide for choosing between line, bar, scatter, histogram, box and pie charts, with the question each one answers and the mistakes to avoid.
  • How to Make a Calibration Curve – Step-by-step calibration curve: plot standards, fit a line, check R² and residuals, then calculate an unknown concentration, LOD and LOQ from real numbers.

Written and reviewed by the Plotiq Editorial Team. Spotted a mistake? Tell us and we will correct it. See About Plotiq for how these guides are made.