Paper Figures

Academic Writing

Expert academic visualization assistant for architecture diagrams, chart type recommendations, and figure/table captions.

Overview

This skill covers four visualization-related tasks for academic papers: creating architecture or method overview diagrams (flat vector style matching DeepMind/OpenAI aesthetics), recommending the best chart type for experiment results from a 19-type academic chart library, and generating figure or table captions that meet top conference standards.

When to Use

• Creating a method overview or architecture diagram for a paper
• Deciding which chart type best presents your experiment results
• Writing concise, standards-compliant figure or table captions
• Getting visualization design specifications (axes, color palette, scale handling)

Key Features

• Architecture diagrams: flat vector illustration style, pastel color palette, pure white background, English text labels only
• 19-type chart library: grouped bars, Pareto fronts, radar charts, confidence-band lines, ROC/PR curves, heatmaps, violin plots, facet grids, and more
• Scale adaptiveness: recommends broken axes, log scales, or normalization based on data distribution
• Caption formatting: Title Case for noun phrases, Sentence case for complete sentences, no redundant "Figure shows" openings
• Dual prompt support: provides both Chinese and English prompts for image generation models

Example Prompts

Draw an architecture diagram for my method based on this abstract and methodology.

What chart type should I use for these SOTA comparison results?

Write a figure caption for this performance comparison plot.

Write a table caption for this ablation study table.

Source

• Skill folder: skills/paper-figures/
• Origin: awesome-ai-research-writing
• Standard: agentskills.io