Chart data generator

The 15 patterns

Each pattern is a formula over the points from left to right, plus a random offset controlled by Noise. Every value is clamped to your minimum and maximum, so a noisy series never leaves the range you set.

Random

Independent values anywhere between minimum and maximum.

Linear

A straight rise from minimum to maximum.

Exponential

A slow start that turns steep toward the end.

Logarithmic

A fast rise that flattens out.

Sine

One full wave around the middle of the range.

Cosine

The same wave, starting at the top.

Trending Up

An upward trend with random steps, like growing sales.

Trending Down

The same with a downward trend.

Seasonal

Two cycles on top of a rising trend.

Stepped

Flat plateaus that jump between levels.

Sparse

Mostly zeros with occasional values, for gaps and rare events.

Random Walk

Each value moves a little from the previous one, like a price.

Bell Curve

One peak in the middle that falls off on both sides.

Sawtooth

Ramps that climb and reset.

Gaussian Noise

Normally distributed values around the middle; Noise sets the spread.

Noise adds a random offset of up to the noise value times the range in either direction: at 0.1 and a range of 0 to 100, each point moves by up to 10. Random, Sparse and Gaussian Noise are random by nature: Random and Sparse ignore Noise, and Gaussian Noise uses it as the spread of the distribution. Random Walk uses a third of the noise, because its steps already vary.

What to test a chart with

Clean demo data hides the problems real data brings. Generate a few uncomfortable sets before you call a chart done:

• Negative values: set the minimum below zero and check the zero line, bar baselines and axis labels.
• Crossing series: three or more series with Random Walk overlap and cross, which shows whether colors, legends and tooltips stay readable.
• Gaps and spikes: Sparse produces mostly zeros with occasional values, the shape of event counts and incomplete sensor data.
• Many points: a few hundred points per series shows whether labels overlap and whether the chart stays fast.
• Long labels: Countries, Cities or your own custom labels test how the axis handles text that does not fit.

The chart and dashboard use case walks through the same checks inside Unity, and Good test data is uncomfortable explains why they matter.

Export formats

The four formats match Mock Magic's export, so files from this page and from the Generator window look the same:

• CSV: one row per point with the columns Label, Value, SecondaryValue, SeriesIndex and SeriesName. Spreadsheets and most chart tools read it directly.
• JSON: one object per series with a name and a data array of label and value pairs. A single series exports as a flat array of points.
• Text: a tab-separated table per series with two decimals, for pasting into documents and issue trackers.
• C#: a string[] of labels and one float[] per series, named after the variable you enter, ready to paste into a test or a MonoBehaviour. Download saves them as MockData.cs inside a static class, so the file compiles on its own, like Save to File in the Generator.

The same series inside Unity

Mock Magic is a free Unity tool that generates these series in the Editor and at runtime. Open Tools > Mock Magic > Generator and switch to Series for the same controls with a live preview, or call the API from code:

using MockMagic;

MockDataPoint[][] series = MockSeries.GenerateMultiSeries(
    SeriesPattern.RandomWalk, seriesCount: 3, pointCount: 24,
    min: 0f, max: 100f, noise: 0.1f, labelPreset: LabelPreset.MonthsShort);

string csv = MockExport.ToCsv(series);

Call Mock.Seed(2026) first to get the same series on every run in Unity. Unity uses a different random number generator than this page, so the shapes match while the individual values differ. Mock Magic can also stream new points live into dashboards, which a static file cannot do: see simulating live data.

To draw the data, Chart Guru includes Mock Magic and fills charts through its Mock Chart Data Source component, and it reads CSV and JSON as well. For single values such as names, e-mail addresses, prices or card numbers, use the fake data generator.

Questions