{"id":5961,"date":"2020-12-09T16:24:02","date_gmt":"2020-12-09T10:54:02","guid":{"rendered":"http:\/\/www.pythonpool.com\/?p=5961"},"modified":"2026-07-13T12:29:19","modified_gmt":"2026-07-13T06:59:19","slug":"matplotlib-log-scale","status":"publish","type":"post","link":"https:\/\/www.pythonpool.com\/matplotlib-log-scale\/","title":{"rendered":"Matplotlib Log Scale: log, symlog, Ticks, and Positive Data"},"content":{"rendered":"

Quick answer:<\/strong> Use ax.set_xscale(‘log’) or ax.set_yscale(‘log’) for positive data spanning orders of magnitude. Choose symlog or another scale when zero or negative values are meaningful, and set locators, formatters, and limits deliberately.<\/p>\n

\"Matplotlib
A log chart clarifies multiplicative structure only when its domain and tick semantics are visible.<\/figcaption><\/figure>\n

A Matplotlib log scale displays axis values by powers instead of equal linear steps. This is useful when data spans several orders of magnitude, such as growth curves, frequencies, execution times, file sizes, population counts, model loss values, or scientific measurements.<\/p>\n

The official Matplotlib documentation for log scale examples<\/a> shows how logarithmic axes change tick spacing. The API references for set_xscale()<\/a>, set_yscale()<\/a>, and loglog()<\/a> cover the main calls.<\/p>\n

Use a log axis when ratios matter more than absolute differences. On a linear y-axis, the change from 10 to 20 looks the same size as the change from 1000 to 1010. On a log y-axis, multiplicative growth is easier to compare because equal visual gaps represent equal ratios.<\/p>\n

The most important rule is that ordinary logarithmic axes need positive values. Zero and negative values cannot be shown on a standard log scale. Filter, transform, or choose a different scale before plotting those values.<\/p>\n

Matplotlib gives you several entry points. semilogx()<\/code> applies a log scale to the x-axis. semilogy()<\/code> applies it to the y-axis. loglog()<\/code> applies it to both axes. You can also create a normal plot and then call set_xscale(\"log\")<\/code> or set_yscale(\"log\")<\/code>.<\/p>\n

A log chart should not be used to hide important differences. It is best when it makes a wide range readable while the caption, axis labels, and surrounding text make the transformed scale obvious to the reader.<\/p>\n

\n
\n

Contents<\/p>\nToggle<\/span><\/path><\/svg><\/svg><\/span><\/span><\/span><\/a><\/span><\/div>\n