Log–log graph: definition, interpretation and applications

Characteristics and interpretation

On a log–log plot equal ratios on an axis correspond to equal distances, so a step from 1 to 10 occupies the same visual space as 10 to 100. Because log transforms are monotonic, ordering is preserved, but zero and negative values cannot be shown directly. If log base changes (for example base‑10 vs natural log), the visual straightness remains but the numerical slope is multiplied by a constant change‑of‑base factor.

Common uses and examples

• Identifying power laws in earth sciences, astronomy, biology, economics and network science.
• Scaling analysis in engineering and materials science (e.g., fracture, fatigue) where exponents describe how one quantity grows with another.
• Empirical data analysis: estimating exponents from the slope of a fitted line after log transformation.

Practical notes and limitations

Fitting lines on log–log axes often uses linear regression on transformed data, but that changes the assumed error model and can bias parameter estimates if original errors are additive. Zeros and negatives must be handled (offsets, censoring or different models). Visual straightness alone does not prove a true power law; statistical tests and alternative models should be considered.

Further reading

For introductions to plotting and interpretation see general resources on logarithmic graphs and scaling laws. A concise overview is available at this reference.