What this workflow helps you do
Prepare a checked scientific plot for its final destination without changing the scientific meaning of the data. This workflow focuses on figure size, axis labels, line and marker visibility, legend placement, annotations and export quality.
Use this workflow when
- Finishing a plot for a journal manuscript, thesis, poster or presentation.
- Making multiple figures visually consistent after the analysis is complete.
- Checking readability at final output size rather than only on a large editor canvas.
What you need before you start
- A dataset or plot whose variables, units and analysis results have already been checked.
- The target figure dimensions or approximate layout space.
- Any journal, thesis or presentation requirements for labels, resolution and file format.
Step-by-step workflow
- Open the checked plot and set the intended output dimensions or aspect ratio.
- Confirm axis titles, units, limits and tick readability at the final size.
- Refine line widths, markers, fonts and legend placement so every series remains distinguishable.
- Add only necessary annotations and keep them clear of important data.
- Preview the figure at or near final display size and check for clipped labels or excessive visual density.
- Export in the required format and resolution, then inspect the exported file before submission.
What to review in the result
- Text remains legible after the figure is inserted at its intended size.
- Colors, line styles and markers remain distinguishable in the destination medium.
- Annotations identify reviewed results and do not imply unsupported conclusions.
- The exported figure contains the same data and processing state that was approved during analysis.
Before you report or export
- Follow the destination journal or institutional figure requirements.
- Keep a reproducible project or data source alongside the final export.
- Prefer concise labels and captions over adding explanatory text directly on the plot.
Limitations and interpretation
- Publication styling cannot correct an incorrect analysis or mislabeled dataset.
- Raster resolution requirements depend on final physical size, not only the pixel count shown in the editor.
- A visually attractive figure should not hide uncertainty, scatter, excluded data or processing choices that matter scientifically.