paper-reproduce
Systematic methodology for reproducing published academic papers using provided data. Use when the user asks to reproduce, replicate, or verify results from a published paper, including sample selection, descriptive statistics, regression analyses, and generating reproduction reports (Markdown + LaTeX PDF). Covers the full pipeline: data exploration, variable identification/mapping, sample filtering, variable construction, statistical analysis, result comparison, and documentation. Applicable to any observational study, clinical cohort, or survey-based research paper.
npx skills add BioTender-max/awesome-bio-agent-skills --skill paper-reproduce --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# Paper Reproduce — 论文复现方法论 ## 核心原则 1. **先探索后建模** — 永远不要假设变量名和编码,必须从数据中验证 2. **逐步筛选逐步核对** — 每一步样本量都要和论文对比 3. **容忍偏差但记录偏差** — harmonized 数据集和原始数据必然有差异,关键是记录和解释 4. **边干边说** — 每完成一步立即输出中间结果,不要等全部跑完才汇报 ## 复现流程(6 阶段) ### Phase 1: 任务理解 + 数据探索 ``` 输入: 论文 PDF/任务文档 + 数据文件 输出: 变量映射表 (variable_mapping.json) ``` 1. 精读任务文档,提取: - 样本筛选流程(每步的 n 和排除条件) - 因变量、自变量、协变量的定义和编码 - 统计方法(回归类型、标准误类型、标准化方式) - 期望的表格数值(用于验证) 2. 探索数据结构: ```python df = pd.read_stata('data.dta') # 或 read_csv/read_sas print(f"维度: {df.shape}") print(f"变量: {df.columns.tolist()}") # 按前缀分组查看变量 # 检查是否有 wave/time/year 标识 ``` 3. **变量识别四步法**(最关键的步骤): - **精确匹配**: 搜索论文提到的变量名(如 `dc009s1`) - **语义搜索**: 搜索关键词(如 `mem`, `recall`, `orient`) - **范围验证**: 检查候选变量的值域是否匹配论文描述(如 0-10) - **交叉验证**: 用已知关系验证(如 `total = sub1 + sub2 + sub3`) > **经验教训**: harmonized 数据集的变量名与原始问卷变量名通常完全不同。 > 不要假设 `dc009s1` 存在,要搜索语义等价的变量。 ### Phase 2: 变量构建 + 验证 ``` 输入: variable_mapping.json 输出: 构建好变量的 DataFrame ``` 1. 逐个构建变量,**每个变量构建后立即验证**: - 值域是否在期望范围内 - 缺失率是否合理 - 均值/比例是否接近论文报告值 2. **组合变量的验证技巧**: ```python # 验证 total = sub1 + sub2 + ... + subN calc = df['sub1'] + df['sub2'] + df['sub3'] diff = (df['total'] - calc).abs() match_rate = (diff < 0.01).mean() print(f"吻合率: {match_rate:.
- 核心原则
- 复现流程(6 阶段)
- Phase 1: 任务理解 + 数据探索
- Phase 2: 变量构建 + 验证
- Phase 3: 样本筛选
- Phase 4: 统计分析
- Phase 5: 结果对比 + 偏差分析
- Phase 6: 输出文档
- 输出目录结构
- 常见陷阱
- 参考
xelatex -interaction=nonstopmode report.tex # 第一遍 xelatex -interaction=nonstopmode report.tex # 第二遍(交叉引用)
What does the paper-reproduce skill do?
Systematic methodology for reproducing published academic papers using provided data. Use when the user asks to reproduce, replicate, or verify results from a published paper, including sample selection, descriptive statistics, regression analyses, and generating reproduction reports (Markdown + LaTeX PDF). Covers the full pipeline: data exploration, variable identification/mapping, sample filtering, variable construction, statistical analysis, result comparison, and documentation. Applicable to any observational study, clinical cohort, or survey-based research paper.
How do I install it?
Run `npx skills add BioTender-max/awesome-bio-agent-skills --skill paper-reproduce --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From BioTender-max/awesome-bio-agent-skills, a repository with 135 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.
