hang-jin/editaplot

What it solves

EditaPlot is designed to automate the creation of professional, publication-quality scientific figures from experimental data. It solves the problem of tedious manual plotting in software like Origin, while ensuring that the figures remain fully editable (OPJU format) and scientifically accurate. Unlike static templates, it prevents the AI from making unfounded assumptions about data, forcing a human-in-the-loop confirmation of data roles and scientific intent before any rendering occurs.

How it works

The tool operates as a Windows-based local skill for Codex, integrating with a local installation of Origin (2021–2026b). The workflow follows a strict sequence:

1. Data Understanding: It reads CSV, TXT, XLS, or XLSX files and identifies the purpose of each column (e.g., primary evidence, auxiliary data, or data to be kept but not plotted).
2. Recommendation & Confirmation: It suggests 1–3 plot types and color palettes. The user must confirm the scientific purpose and the specific elements to be plotted.
3. Visual Style Adaptation: Users can provide a reference image to extract visual grammar and style (layout, markers, colors) without copying the actual data or text from that image.
4. Automated Rendering: It launches a dedicated Origin instance to generate the plot and exports it as an editable OPJU project, as well as PNG, PDF, and TIF files.
5. Verification: It performs an "object back-read" and visual check to ensure axes, fonts, and data mapping are correct before finalizing.

Who it’s for

Researchers and scientists in fields such as materials science (XPS, XRD, XAS, FTIR), chemistry, medicine, and deep learning (SHAP analysis, ROC curves) who use Origin for their final figures.

Highlights

• Fully Editable Outputs: Produces native Origin project files (OPJU) rather than just static images.
• Scientific Guardrails: Refuses to automatically smooth data, fill missing values, or infer conclusions; requires explicit user confirmation for data roles.
• ** uma wide range of specialized templates**: Supports over 40 plotting routes, including complex materials spectroscopy (XPS, XRD Rietveld), medical SHAP summaries, and circular weighted networks.
• Reference Image Guidance: Can mimic the visual style of a reference figure while maintaining the integrity of the user's own experimental data.
• Strict Local Execution: Operates on Windows 10/11 x64, ensuring data stays local and does not upload to a cloud server during the rendering process.