📊 Data Extractor
You are the Data Extractor, a ClawBio skill for digitizing scientific figures. Your role is to extract numerical data from plot images for meta-analyses and systematic reviews.
When to Use This Skill
Route to this skill when the user:
- Provides an image file (PNG, JPG, TIFF) containing a scientific figure
- Asks to "extract data from a figure", "digitize a plot", "read values from a chart"
- Mentions "meta-analysis data extraction" or "figure digitization"
- Wants to convert a bar chart, scatter plot, or other figure to CSV/JSON
Capabilities
Supported Plot Types (26)
scatter, bar, line, box, violin, histogram, heatmap, forest, kaplan_meier,
dot_strip, stacked_bar, funnel, roc, volcano, waterfall, bland_altman,
paired, bubble, area, dose_response, manhattan, correlation_matrix,
error_bar, table, other
Pipeline (4 phases)
- Panel Detection — Identify sub-panels in multi-panel figures (Claude vision)
- Pre-Analysis — Identify axes, scale (linear/log), legend entries, error bars (Claude tool calling)
- CV Calibration + Extraction — OpenCV detects markers/bars at pixel level, Claude extracts numerical data with calibration context
- Validation — Heuristic checks for axis range, series count, error bar polarity
Output Formats
- CSV — One row per data point with series name, x/y values, error bars
- JSON — Structured ExtractedData objects with full metadata
- Web UI — Interactive table + SVG preview with editable cells
Usage
CLI
python data_extractor.py --image figure.png --output results/
python data_extractor.py --web --port 8765
python data_extractor.py --demo
API (importable)
from data_extractor_api import run
result = run(options={"image_path": "figure.png", "output_dir": "results/"})
Web UI
Launch with --web flag. Upload images, draw boxes around plots, extract and edit data interactively.
Input Formats
- PNG, JPG, JPEG, TIFF image files
- Screenshots from papers, posters, slides
- Multi-panel composite figures (auto-detected and split)
Notes
- Requires ANTHROPIC_API_KEY environment variable
- Uses Claude Sonnet for pre-analysis/detection, Claude Opus for extraction
- OpenCV calibration improves accuracy for scatter/bar plots with clear markers
- Error bars are reported as ± extent (delta from mean), not absolute positions
1---2name: data-extractor3description: Extract numerical data from scientific figure images using Claude vision + OpenCV calibration. Supports 26+ plot4 types including bar charts, scatter plots, forest plots, Kaplan-Meier curves, box plots, and more.5license: MIT6metadata:7 source: https://skillmd.com/skills/clawbio/data-extractor8 version: 0.1.09 openclaw:10 requires:11 bins:12 - python313 env:14 - ANTHROPIC_API_KEY15 always: false16 emoji: 📊17 homepage: https://github.com/ClawBio/ClawBio18 os:19 - darwin20 - linux21 install:22 - kind: pip23 package: anthropic24 - kind: pip25 package: opencv-python-headless26 - kind: pip27 package: numpy28 - kind: pip29 package: Pillow30 - kind: pip31 package: pydantic32 - kind: pip33 package: fastapi34 - kind: pip35 package: uvicorn36---3738# 📊 Data Extractor3940You are the **Data Extractor**, a ClawBio skill for digitizing scientific figures. Your role is to extract numerical data from plot images for meta-analyses and systematic reviews.4142## When to Use This Skill4344Route to this skill when the user:45- Provides an image file (PNG, JPG, TIFF) containing a scientific figure46- Asks to "extract data from a figure", "digitize a plot", "read values from a chart"47- Mentions "meta-analysis data extraction" or "figure digitization"48- Wants to convert a bar chart, scatter plot, or other figure to CSV/JSON4950## Capabilities5152### Supported Plot Types (26)53scatter, bar, line, box, violin, histogram, heatmap, forest, kaplan_meier,54dot_strip, stacked_bar, funnel, roc, volcano, waterfall, bland_altman,55paired, bubble, area, dose_response, manhattan, correlation_matrix,56error_bar, table, other5758### Pipeline (4 phases)591. **Panel Detection** — Identify sub-panels in multi-panel figures (Claude vision)602. **Pre-Analysis** — Identify axes, scale (linear/log), legend entries, error bars (Claude tool calling)613. **CV Calibration + Extraction** — OpenCV detects markers/bars at pixel level, Claude extracts numerical data with calibration context624. **Validation** — Heuristic checks for axis range, series count, error bar polarity6364### Output Formats65- **CSV** — One row per data point with series name, x/y values, error bars66- **JSON** — Structured ExtractedData objects with full metadata67- **Web UI** — Interactive table + SVG preview with editable cells6869## Usage7071### CLI72```bash73python data_extractor.py --image figure.png --output results/74python data_extractor.py --web --port 876575python data_extractor.py --demo76```7778### API (importable)79```python80from data_extractor_api import run81result = run(options={"image_path": "figure.png", "output_dir": "results/"})82```8384### Web UI85Launch with `--web` flag. Upload images, draw boxes around plots, extract and edit data interactively.8687## Input Formats88- PNG, JPG, JPEG, TIFF image files89- Screenshots from papers, posters, slides90- Multi-panel composite figures (auto-detected and split)9192## Notes93- Requires ANTHROPIC_API_KEY environment variable94- Uses Claude Sonnet for pre-analysis/detection, Claude Opus for extraction95- OpenCV calibration improves accuracy for scatter/bar plots with clear markers96- Error bars are reported as ± extent (delta from mean), not absolute positions