SCIENTIFIC GRAPH STUDIO / GUIDE

A practical route
from data to figure.

A high-level workflow guide. Menu names and availability can vary by build; ask support for version-specific details.

01 / DATA

Prepare before you plot.

Use a clearly structured CSV or Excel worksheet. Keep measured quantities in separate columns and retain an unmodified copy of the source data.

Check the selected sheet, header row, units, and numerical interpretation. Decimal separators, blank cells, and values stored as text can affect which points are available.

Before analysis: missing values and excluded points should be a conscious decision, not an unnoticed import side effect.
02 / FIGURE

Map variables, then choose the presentation.

Assign the relevant worksheet and X/Y columns. Choose a plot type that reflects the data structure. Check axes, scales, legends, and panel organization.

When using multiple axes, make units and assignments unmistakable. For error bars, state whether the values represent standard deviation, standard error, confidence intervals, or another defined quantity.

03 / INTERPRETATION

A fit is not independent validation.

Choose a fitting relationship for a stated purpose. Record the interval, parameter meaning, units, and preprocessing. Inspect the data and residual behavior, not only a summary fit statistic.

Do not silently extrapolate beyond the data range or the physical assumptions that justify a model. Keep estimates visually distinct from measurements.

04 / OUTPUT

Review the file people will actually see.

Set the intended physical figure dimensions. Vector formats can preserve editable text and paths, while raster outputs require resolution suited to their final size.

Check the exported file for clipped labels, missing characters, legend readability, and altered dimensions. Publication templates are starting points, not acceptance guarantees.

Open the figure checklist ↗
HELP

When something looks wrong.

Start with the source worksheet, selected columns, axis range, and active filters. A plotting issue may be a data-interpretation issue rather than a rendering issue.

For support, include the version, operating system, expected result, actual result, and a minimal anonymized example. Do not send restricted or proprietary data.

Contact software support