# Asreview Screening

Screen papers for systematic reviews using ASReview active learning.

- Repository: https://github.com/beita6969/ScienceClaw (1024 stars)
- Category: Engineering
- License: MIT
- Last updated: 2026-10-31
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add beita6969/ScienceClaw --skill asreview-screening -g -a claude-code -y`
Codex: `npx skills add beita6969/ScienceClaw --skill asreview-screening -g -a codex -y`
Cursor: `npx skills add beita6969/ScienceClaw --skill asreview-screening -g -a cursor -y`
Gemini CLI: `npx skills add beita6969/ScienceClaw --skill asreview-screening -g -a gemini-cli -y`

## Description

Screen papers for systematic reviews using ASReview active learning. Use when: user has a large set of papers to screen for inclusion/exclusion, wants to prioritize relevant papers, or needs to reduce manual screening workload. NOT for: searching papers (use literature-search) or meta-analysis (use meta-analysis).

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# ASReview Screening

Use active learning to prioritize and screen papers for systematic reviews, reducing manual workload by up to 95%. ASReview uses machine learning to learn from your screening decisions and prioritize the most likely relevant papers.

## When to Use

- "I have 500 papers to screen for my systematic review"
- "Help me prioritize papers for inclusion"
- "Set up active learning screening for my review"
- "How many papers do I need to screen manually?"

## When NOT to Use

- Searching for papers (use literature-search)
- Performing meta-analysis (use meta-analysis)
- Writing the review (use systematic-review + paper-writing)
- Small sets (< 50 papers), manual screening is faster

## Setup

### Install ASReview

```bash
pip install asreview asreview-insights asreview-datatools
```

### Prepare Input Data

ASReview accepts RIS, CSV, TSV, or Excel files with at minimum:
- `title`: Paper title
- `abstract`: Paper abstract

Optional but recommended:
- `doi`, `authors`, `year`, `keywords`, `label` (if some are pre-labeled)

### Export from Search Results

```python
# Convert Semantic Scholar / OpenAlex results to ASReview format
import csv

def export_for_asreview(papers: list[dict], output_path: str):
    """Export papers to CSV for ASReview."""
    with open(output_path, 'w', newline='', encoding='utf-8') as f:
        writer = csv.DictWriter(f, fieldnames=[
            'title', 'abstract', 'authors', 'year', 'doi', 'keywords'
        ])
        writer.writeheader()
        for p in papers:
            writer.writerow({
                'title': p.get('title', ''),
                'abstract': p.get('abstract', ''),
                'authors': '; '.join(a.get('name', '') for a in p.get('authors', [])),
                'year': p.get('year', ''),
                'doi': p.get('externalIds', {}).get('DOI', ''),
                'keywords': '; '.join(p.get('fieldsOfStudy', []))
            })
    print(f"Exported {len(papers)} papers to {output_path}")
```

The rest: https://github.com/beita6969/ScienceClaw/blob/HEAD/skills/asreview-screening/SKILL.md