Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Research posters in LaTeX using beamerposter, tikzposter, or baposter. Layout, typography, color schemes, figure integration, accessibility, and QA for conferences. Includes templates. For figure generation use matplotlib-scientific-plotting or plotly-interactive-plots.
.claude/skills/jaechang-hits-latex-research-posters/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 51% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 122% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 169% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 106% | 0% |
| case-01 | ✓→✗ | ▼ Worse | 62% | 0% |
Research posters are a critical medium for scientific communication at conferences, symposia, and academic events. This knowhow covers end-to-end poster creation in LaTeX: package selection, layout design, typography, color schemes, figure integration, accessibility, compilation, and quality control for print and digital display.
Three major packages, each with distinct strengths:
| Package | Architecture | Best For | Learning Curve | |---------|-------------|----------|---------------| | beamerposter | Beamer extension | Traditional academic posters, institutional branding | Low (if you know Beamer) | | tikzposter | TikZ-based blocks | Modern colorful designs, custom graphics | Medium | | baposter | Box-based grid | Multi-column layouts, consistent spacing | Low |
beamerposter: Uses Beamer's \begin{block} syntax. Themes and color schemes from Beamer carry over. Best when your institution already has a Beamer theme.
tikzposter: Declarative block placement with \block{Title}{Content}. Built-in color styles (Denmark, Germany, etc.) and layout themes (Rays, Wave). Most flexible for custom designs.
baposter: Defines named boxes in a grid. Automatic positioning and spacing. Best when you want structured, uniform column layouts without manual placement.
| Standard | Size | Region | Use | |----------|------|--------|-----| | A0 | 841 × 1189 mm (33.1 × 46.8 in) | Europe | Most common academic standard | | A1 | 594 × 841 mm (23.4 × 33.1 in) | Europe | Smaller venues | | 36 × 48 in | 914 × 1219 mm | North America | Standard US conference size | | 42 × 56 in | 1067 × 1422 mm | North America | Large format | | 48 × 72 in | 1219 × 1829 mm | North America | Extra large |
Orientation: Portrait (vertical) is most common and traditional. Landscape (horizontal) works better for timelines, wide figures, or side-by-side comparisons.
| Element | Size Range | Purpose | |---------|-----------|---------| | Title | 72–120 pt | Readable from 15+ feet | | Section headers | 48–72 pt | Readable from 8–10 feet | | Body text | 24–36 pt | Readable from 4–6 feet | | Captions | 20–28 pt | Readable from 3 feet |
Start: What is your design priority?
├── Institutional branding / existing Beamer theme?
│ └── YES → beamerposter
├── Modern, colorful, custom TikZ graphics?
│ └── YES → tikzposter
├── Structured multi-column grid with minimal setup?
│ └── YES → baposter
└── Not sure
└── tikzposter (most flexible default)| Poster Content | Columns | Layout Strategy | |---------------|---------|-----------------| | Few key results, large figures | 2 columns | Wide figure panels, brief text | | Balanced text and figures | 3 columns | Standard academic layout | | Data-heavy with many small figures | 4 columns | Compact grid, small text | | Narrative flow / timeline | 2 columns landscape | Left-to-right story |
| Poster Type | Text % | Visual % | Word Count | |-------------|--------|----------|------------| | Experimental research | 40% | 60% | 400–600 | | Computational/modeling | 50% | 50% | 500–700 | | Review/survey | 55% | 45% | 600–800 | | Method paper | 35% | 65% | 300–500 |
pdfinfo poster.pdf | grep "Page size" after compilation. A0 should show ~2384 × 3370 points.pdffonts poster.pdf — every font should show "yes" in the "emb" column. Non-embedded fonts may render differently on the printer's system.pdfimages -list poster.pdf to verify.pdflatex (auto-embeds), or use -dEmbedAllFonts=true. Verify with pdffonts.\includegraphics with consistent sizing and clear captionspdflatex poster.tex (or lualatex for better font support)pdfinfo poster.pdf | grep "Page size" — must match requirements exactlypdffonts poster.pdf — all "yes" in "emb" columnpdfimages -list poster.pdf — all ≥300 DPI[LastName]_[Conference]_Poster.pdfbash# Basic compilation pdflatex poster.tex # With bibliography pdflatex poster.tex && bibtex poster && pdflatex poster.tex && pdflatex poster.tex # Better font support lualatex poster.tex # or xelatex poster.tex
bash# Verify page dimensions pdfinfo poster.pdf | grep "Page size" # Check font embedding pdffonts poster.pdf # Check image resolution pdfimages -list poster.pdf # Check for compilation warnings grep -i "warning\|error\|overfull\|underfull" poster.log # Compress for email (keeps print quality) gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 \ -dPDFSETTINGS=/printer -dNOPAUSE -dQUIET -dBATCH \ -sOutputFile=poster_compressed.pdf poster.pdf
bash# TeX Live (Linux/Mac) tlmgr install beamerposter tikzposter baposter qrcode subcaption tcolorbox # MiKTeX (Windows) — packages auto-install on first use
Templates are provided in the assets/ subdirectory:
| File | Package | Description | |------|---------|-------------| | beamerposter_template.tex | beamerposter | Traditional 3-column portrait layout (A0) with institutional header | | tikzposter_template.tex | tikzposter | Modern block-based design with Denmark color style | | baposter_template.tex | baposter | Structured 3-column portrait grid with automatic spacing |
Each template is a complete, compilable LaTeX document with placeholder content. Customize: (1) replace placeholder text and figures, (2) adjust color scheme, (3) update page size if needed.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→fail | 25,422 | 23,208 | -9% | 1 | 1 | 0% | 4,555 | 7,394 | +62% | 0 | 0 | — |
case-02 | fail→pass | 32,906 | 30,949 | -6% | 1 | 1 | 0% | 6,215 | 9,400 | +51% | 0 | 0 | — |
case-03 | pass→pass | 11,744 | 6,579 | -44% | 1 | 1 | 0% | 2,105 | 4,673 | +122% | 0 | 0 | — |
case-04 | fail→pass | 12,092 | 6,847 | -43% | 1 | 1 | 0% | 2,081 | 4,630 | +122% | 0 | 0 | — |
case-05 | pass→pass | 10,319 | 5,070 | -51% | 1 | 1 | 0% | 1,745 | 4,413 | +153% | 0 | 0 | — |
case-06 | fail→pass | 10,716 | 7,562 | -29% | 1 | 1 | 0% | 1,748 | 4,710 | +169% | 0 | 0 | — |
case-07 | pass→pass | 10,766 | 5,373 | -50% | 1 | 1 | 0% | 1,790 | 4,428 | +147% | 0 | 0 | — |
case-08 | pass→pass | 9,040 | 5,412 | -40% | 1 | 1 | 0% | 1,577 | 4,406 | +179% | 0 | 0 | — |
case-09 | pass→pass | 7,584 | 4,149 | -45% | 1 | 1 | 0% | 1,384 | 4,147 | +200% | 0 | 0 | — |
case-10 | pass→pass | 7,745 | 4,008 | -48% | 1 | 1 | 0% | 1,499 | 4,180 | +179% | 0 | 0 | — |
case-11 | pass→pass | 14,306 | 14,620 | +2% | 1 | 1 | 0% | 2,477 | 6,195 | +150% | 0 | 0 | — |
case-12 | pass→pass | 8,868 | 2,523 | -72% | 1 | 1 | 0% | 1,609 | 3,861 | +140% | 0 | 0 | — |
case-13 | pass→pass | 9,938 | 1,960 | -80% | 1 | 1 | 0% | 1,626 | 3,810 | +134% | 0 | 0 | — |
case-14 | pass→pass | 10,322 | 7,449 | -28% | 1 | 1 | 0% | 1,800 | 4,803 | +167% | 0 | 0 | — |
case-15 | pass→pass | 7,036 | 10,843 | +54% | 1 | 1 | 0% | 1,184 | 3,828 | +223% | 0 | 0 | — |
case-16 | pass→pass | 11,168 | 5,132 | -54% | 1 | 1 | 0% | 1,869 | 4,309 | +131% | 0 | 0 | — |
case-17 | fail→pass | 13,428 | 6,305 | -53% | 1 | 1 | 0% | 2,146 | 4,415 | +106% | 0 | 0 | — |
case-18 | pass→pass | 10,815 | 11,382 | +5% | 1 | 1 | 0% | 1,815 | 5,260 | +190% | 0 | 0 | — |
case-19 | pass→pass | 8,855 | 4,237 | -52% | 1 | 1 | 0% | 1,578 | 4,197 | +166% | 0 | 0 | — |
case-20 | pass→pass | 16,052 | 14,267 | -11% | 1 | 1 | 0% | 2,772 | 6,158 | +122% | 0 | 0 | — |
case-21 | pass→pass | 21,929 | 27,771 | +27% | 1 | 1 | 0% | 4,257 | 8,729 | +105% | 0 | 0 | — |
case-22 | pass→pass | 15,943 | 19,800 | +24% | 1 | 1 | 0% | 3,389 | 5,700 | +68% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.