Logarithmic Dashboard Axes: When Equal Steps Mean Ratios, Not Equal Amounts

A logarithmic axis can help a reporting screen compare positive values spanning very different magnitudes. But it changes the meaning of visual distance: equal steps represent equal multiplication factors, not equal amounts added. Commission it only when that comparison serves the reader's question, and make the scale explicit.

This is a guide to specifying a dashboard, not a recommendation to transform every chart with a large maximum. An attractive fit inside the available space does not establish that the result answers the right business question.

Decide whether the question concerns amounts or proportions

Imagine a hypothetical operational report with measurements of 1, 10, 100 and 1,000 in the same unit. On a base-ten logarithmic axis, successive values occupy equal steps. Each is ten times the preceding value. Their absolute differences are not equal.

If the reader wants to identify the largest additional workload, a linear view of the original amounts may be more useful. If the question concerns multiplicative differences across a wide range, a logarithmic view may be worth commissioning. Write that purpose into the report definition before selecting the chart option.

The official D3 documentation for logarithmic scales describes applying a logarithmic transform to the input before mapping it into the output range. It gives the mapping as a scaled and shifted logarithm. The equal-ratio interpretation follows from that transform; it does not mean the business measurements themselves have been replaced with different facts.

Keep the original values available. A user should be able to inspect a point and understand its actual measurement and unit, not merely see a transformed coordinate whose business meaning is unexplained.

Do not confuse a logarithmic axis with a percentage metric

A chart can show original counts on a logarithmic axis without turning those counts into percentage changes. Conversely, a report of percentage changes can still use a linear axis. The measure and the position encoding are separate decisions.

My recommendation is to name both in the brief. For example, specify that the plotted measure is an observed duration in a stated unit, while the axis is logarithmic to compare relative size. Do not label the result “growth” merely because the display uses logarithms. Growth requires an actual comparison across observations or periods with a defined meaning.

Also avoid presenting a smoother-looking line as evidence that a process has become more stable. The scale can alter the visual prominence of large absolute changes. It does not improve the underlying operation or validate a prediction.

Decide what happens to zero before building the chart

Zero is not a minor formatting issue. D3 warns that a log-scale domain must be strictly positive or strictly negative and must not include or cross zero. For a positive domain, supplying a negative value has undefined behaviour; the converse applies to a negative domain.

That does not mean negative business values are universally invalid. D3 supports an entirely negative domain through an internal sign treatment. It does mean that a series mixing positive and negative values cannot simply be passed to this ordinary logarithmic scale as though it were linear.

For a positive-measure report, agree how the interface handles a genuine zero, an unexpected negative measurement, an absent value and a failed data request. These states have different meanings. Do not replace a real zero with an arbitrary small positive number just to make it plot. Do not silently remove nonpositive observations and leave the reader believing the chart is complete.

Possible project choices include using a linear view for the relevant selection or clearly separating observations the chosen scale cannot represent. A different transform would require its own reviewed interpretation. This guide does not establish that any alternative is appropriate for your data.

Label the scale where readers will encounter it

I recommend a visible “logarithmic scale” label and an explanation that equal distances indicate equal factors. Retain readable tick values in the original unit. A tooltip is useful additional detail, but should not be the only place a reader can discover that the axis is nonlinear.

D3 distinguishes tick generation from tick-label formatting. Its documentation notes that the formatter can leave some labels empty when there are too many ticks, while the ticks themselves remain. A narrow screen can therefore deserve a separate inspection: the presence of gridlines does not guarantee that the remaining labels explain the scale clearly.

The same documentation says that changing the logarithmic base affects which ticks are chosen, not the scale's encoding. Do not offer a base selector as if it changes the recorded measurements or reveals a new business conclusion. If a base choice is exposed, explain its limited purpose and test the resulting labels.

Preserve the comparison when filters or exports change

Decide whether a filter keeps the same scale type and range or deliberately changes them. An unannounced switch from linear to logarithmic can make successive screenshots appear comparable when they encode distance differently.

For a proposed comparison view, keep scale type, units and population scope visible. A permitted export should retain the scale explanation and original values needed to interpret it. A saved chart should not rely on an interactive control that disappears outside the application.

These are commissioning recommendations, not capabilities demonstrated in a particular reporting library. A suitable implementation also needs an accessible explanation and readable data, but this article's decision is specifically about multiplicative distance and the scale's input constraints, not a complete chart-accessibility assessment.

Include the difficult values in acceptance work

Add explicit scale cases to the acceptance criteria for your web app:

Positive measurements related by the same factor occupy the agreed equal distances, with the original values still inspectable.
A measured zero is handled according to the declared policy, not silently changed to a positive observation.
Negative, mixed-sign, missing and unavailable inputs receive distinct, truthful treatment appropriate to the commissioned report.
Tick labels, units and the scale explanation remain understandable on a narrow screen.
Filters and exports preserve or explicitly disclose any change in the comparison scale.
The readable values and visual points refer to the same selected records.

Use harmless synthetic observations for these proposed tests. No dashboard or production dataset was executed for this article. Bring the comparison question, the expected value range and examples of exceptional values to a reporting-app discussion. Those details are more useful than asking for a chart that simply looks less crowded.

Source and limits

The D3 logarithmic-scale documentation was read on 26 October 2026. It supports the transform definition, domain and sign restrictions, base behaviour and distinction between ticks and displayed tick labels. Business-purpose selection, input-state handling, export context and acceptance requirements above are practical product guidance. They are not a statistical certification, legal conclusion or claim that a particular app has passed testing.