graph-interpretation

What This Skill Does

The graph-interpretation skill acts as an advanced visual reasoning engine specifically calibrated for scientific and clinical documentation. It processes complex visual data such as Kaplan-Meier curves, forest plots, ROC curves, and heatmaps, transforming them into structured, rigorous analytical text. By leveraging statistical extraction, the agent identifies core clinical metrics—like hazard ratios, p-values, confidence intervals, and effect sizes—ensuring that the generated insights adhere to academic reporting standards. It translates raw pixel data into accurate narrative descriptions suitable for peer-reviewed papers, clinical trial reports, or public-facing summaries.

Installation

To integrate this skill into your environment, run the following command in your terminal: clawhub install openclaw/skills/skills/aipoch-ai/graph-interpretation Ensure you have the OpenClaw environment initialized and permissions granted for the required file access for image processing.

Use Cases

• Academic Publishing: Automatically drafting figure captions and results sections for manuscripts based on study images.
• Clinical Reporting: Rapidly interpreting patient-centric trial data for presentations to internal medical committees.
• Grant Writing: Summarizing complex data distributions from pilot studies to illustrate project feasibility and impact.
• Policy Briefings: Explaining the significance of clinical research outcomes to non-technical stakeholders or policymakers.

Example Prompts

1. "Analyze this Kaplan-Meier plot and generate a summary statement for the results section of a phase 3 oncology manuscript. Focus on the hazard ratio and p-value."
2. "I am presenting this forest plot to a group of clinicians. Please rewrite the key findings in simple language, explaining the significance of the 95% confidence interval."
3. "Look at this volcano plot from our omics dataset and identify the top three genes that pass the FDR threshold. Write a short caption describing the differential expression pattern."

Tips & Limitations

• Image Quality: Ensure that images provided are high-resolution, as low-quality charts may lead to inaccuracies in numerical data extraction. Always verify critical p-values against original source spreadsheets.
• Context is Key: Providing specific context (e.g., 'oncology_phase3_trial') significantly improves the agent's ability to interpret ambiguous axes or labels.
• Verification: While the skill provides high-level analytical insight, it should be treated as an assistant for drafting; clinical findings must always be reviewed by a human researcher or principal investigator to ensure 100% precision in final reports.