Why Matplotlib Is Still the Standard
Despite the rise of Seaborn, Plotly, and Altair, Matplotlib remains the tool you need to know for academic publishing. Every journal has figure requirements (DPI, font size, line weight, column width) and Matplotlib is the most direct way to meet them precisely.
Seaborn and Plotly are excellent for exploratory analysis. Matplotlib is what you use when you need the final, camera-ready figure.
Publication-Ready in 5 Steps
- Set figure size to match your journal’s column width — single-column figures are typically 3.5 inches; double-column is 7 inches. Set this in inches, not pixels.
- Use the right DPI — journals typically require 300 DPI for raster formats. Use
plt.savefig('figure.pdf')for vector output when possible. - Match font sizes — captions are typically 8–10pt. Set
rcParams['font.size'] = 9globally. - Use consistent colours — define a palette at the top of your script and reuse it across all figures in a paper.
- Export as PDF or EPS — vector formats scale without pixelation, are universally accepted, and can be opened in Inkscape for final touch-ups.
Animations for Research
The animation tutorial shows how to create FuncAnimation visualisations — useful for:
- Showing how a model converges over training steps
- Visualising particle simulations or fluid dynamics
- Conference presentations where a moving plot communicates better than a static one
The Matplotlib + Inkscape Pipeline
My recommended workflow:
- Generate the base figure in Matplotlib (axes, data, labels)
- Export as PDF or SVG
- Open in Inkscape to fine-tune arrow positions, add annotations, or adjust text
- Export to final format for submission
This keeps the data layer (Python) separate from the design layer (Inkscape) — making it easy to update data without redoing the design.