Releases: SuperJayLiu/AI-for-Economics-and-Finance-Research
Release list
AI_EFR_v1.0.1
AI for Economics and Finance Research — v1.0.1
A maintenance-and-expansion release on top of v1.0.0. The handbook is now also a citable PDF working paper, the entire repository can be installed as a single AI skill, every skill has been upgraded to a more disciplined operating protocol, and the data and presentation guidance has grown.
Same simple loop as before: read → copy a skill → use it on a real task → verify the output → record what changed.
What's new in v1.0.1
- 📄 PDF working-paper edition (SSRN). The full handbook is now available as a single, readable, citable working paper, in addition to the GitHub version.
- 🧩 Use the whole repository as one AI skill. A new top-level
SKILL.mdrouter and a self-containedskill-pack/folder (a portableSKILL.mdplus 14 short reference files) let you paste one file into a ChatGPT or Claude project and have the assistant route each task through the handbook's sections and rules. - 🧠 Every skill upgraded with an "import/use-as-skill" protocol. Skills now collect only the missing inputs (and ask up to five clarifying questions), return a "Proposed structure and assumptions" before long or file-producing outputs, double-check for unsupported claims and privacy risks, and always end with What I produced / What I did not change / What you must verify / Questions for you.
- 🌍 Expanded data guidance. New coverage of survey, household-finance, and Nordic register data — ECB HFCS, Eurostat public microdata, SHARE, European Social Survey, LIS/LWS, and the Nordic registers (Statistics Denmark, Sweden's MONA, Norway / microdata.no / Sikt, Finland's FIONA, Iceland) — framed as access maps, not upload permission, alongside expanded aggregate and market-data notes.
- 🎤 Expanded presentation tools. New copy-ready skills for full-materials structure intake, interactive single-file HTML slides, and LaTeX/Beamer slides, plus a new "build a simple academic website" skill (GitHub Pages or Google Sites).
- 🗂️ Release housekeeping. Updated
CITATION.cff, SSRN link, changelog, and an acknowledgement to the "AI-self-help group" at USI Lugano and to scholars sharing their work openly.
What's included (full repository)
- 01 · Start Here — The Handbook. Mental model, maturity ladder, what AI is and isn't good at, the econ/finance workflow, and data, GitHub, and project safety.
- 02 · Copy and Use — Skills and Templates. A library of 110+ copy-ready blocks across research ideas, literature review, paper drafting and citations, economics and finance empirical methods, causal inference (DiD, IV, RD, synthetic control, panel FE, time series), coding and debugging (Stata/R/Python), data cleaning and merging, text-as-data / LLM measurement, theory and proofs, structural and welfare, research-question framing, referee reports, presentations and websites, and verification and disclosure.
- 03 · Set Up Agents and Workflows. 6 Git-safe workflows: clean-up and Git setup, one-paper-one-repo, replication packages, update digests, parallel agents and worktrees, and GitHub review/publish.
- 04 · Examples, Diagrams, and Failure Cases. Worked examples, a research-workflow diagram, and a failure-case library.
- 05 · Check Sources and Official Docs. A source-to-repo map, a source-use policy, the expanded data resources above, and privacy-safe ways to stay updated.
- 06 · Teach, Practice, and Share Slides. Workshop outlines (including a two-hour session with slides), RA onboarding checklists, and presentation-practice activities.
SKILL.md+skill-pack/. Install the whole handbook as one AI skill (new in v1.0.1).
Who it's for
Students, research assistants, PhD students, and faculty in economics and finance who want the genuine productivity gains of AI without surrendering the judgment, rigor, and integrity that good scholarship depends on.
Core principle
AI can automate labor, but not scholarly responsibility. Treat every AI output as a draft, checklist, critique, or coding aid — and verify claims, citations, coefficients, code, and equations against original sources before using them. Always follow your institution, funder, journal, and data-provider rules; where they are stricter, follow the stricter rule.
Languages
Fully bilingual — English and 中文 (Simplified Chinese).
License and citation
Released under the MIT License. If this repository supports your research or teaching, please cite it using the included CITATION.cff, or cite the SSRN working paper.
Connect
- LinkedIn: Chaojie (Jay) Liu
- 小红书 (RED): SuperJay
Feedback and contributions
Questions, corrections, or new-skill suggestions are welcome: jay.liu@bristol.ac.uk — subject line [AI Econ Finance Handbook] Question or correction.
高效地将 AI 应用于经济与金融研究 — v1.0.1
在 v1.0.0 基础上的一次维护与扩展更新。本手册现在还提供可引用的 PDF 工作论文版本，整个仓库可作为单一 AI skill 安装使用，所有 skill 都升级为更规范的操作协议，数据与演示部分也进一步扩充。
依然是同一个简单的循环：阅读 → 复制一个 skill → 在真实任务上使用 → 核查输出 → 记录改动。
v1.0.1 新增内容
- 📄 PDF 工作论文版本（SSRN）。 除 GitHub 版本外，完整手册现以单一、可阅读、可引用的工作论文形式提供。
- 🧩 将整个仓库作为一个 AI skill 使用。 新增顶层
SKILL.md路由文件，以及一个自包含的skill-pack/目录（一个可移植的SKILL.md加 14 个简短参考文件）——把一个文件粘贴进 ChatGPT 或 Claude 项目，助手即可按手册的章节与规则为每个任务做路由。 - 🧠 所有 skill 升级为"导入即用"操作协议。 skill 现在只收集缺失的输入（并最多提出五个澄清问题），在生成长篇或会产出文件的输出前先返回"建议结构与假设（Proposed structure and assumptions）"，并在结束前自查是否存在无依据的论断与隐私风险，最后统一以 我产出了什么 / 我没有改动什么 / 你必须核查什么 / 给你的问题 收尾。
- 🌍 扩充数据指引。 新增对调查数据、家庭金融与北欧登记数据的覆盖——ECB HFCS、Eurostat 公开微观数据、SHARE、European Social Survey、LIS/LWS，以及北欧登记数据（丹麦统计局、瑞典 MONA、挪威 / microdata.no / Sikt、芬兰 FIONA、冰岛）——并将其定位为访问地图，而非上传许可；同时扩充了宏观与市场数据来源说明。
- 🎤 扩充演示工具。 新增可直接使用的 skill：完整材料的结构确认（intake）、交互式单文件 HTML 幻灯片、LaTeX/Beamer 幻灯片，以及一个新的**"搭建简单的学术个人网站"** skill（GitHub Pages 或 Google Sites）。
- 🗂️ 发布维护。 更新了
CITATION.cff、SSRN 链接、更新日志，并致谢卢加诺大学（USI Lugano）的"AI 自助小组"以及公开分享其工作的各位学者。
包含内容（完整仓库）
- 01 · 从这里开始 — 手册。 思维模型、成熟度阶梯、AI 擅长与不擅长的事、经济/金融研究工作流，以及数据、GitHub 与项目安全。
- 02 · 复制与使用 — Skills 与模板。 110+ 个可直接复制使用的模块，涵盖研究想法、文献综述、论文写作与引用、经济与金融实证方法、因果推断（DiD、IV、RD、合成控制、面板固定效应、时间序列）、编程与调试（Stata/R/Python）、数据清理与合并、text-as-data / LLM 测量、理论与证明、结构与福利、研究问题构建、审稿报告、演示与网站，以及核查与披露。
- 03 · 搭建 Agents 与工作流。 6 个 Git 安全工作流：清理项目与 Git 设置、one-paper-one-repo、复现包、更新摘要、并行 agents 与 worktrees，以及 GitHub review/publish。
- 04 · 实例、图示与失败案例。 实例、研究工作流图示，以及失败案例库。
- 05 · 核查来源与官方文档。 来源到仓库的映射、来源使用政策、上述扩充的数据资源，以及保护隐私的更新方式。
- 06 · 教学、练习与分享幻灯片。 工作坊大纲（含两小时课程与幻灯片）、RA 入门清单，以及演讲练习活动。
SKILL.md与skill-pack/。 将整个手册作为一个 AI skill 安装使用（v1.0.1 新增）。
适合谁
经济学与金融学领域的学生、研究助理、博士生与教师——希望获得 AI 带来的真实效率提升，同时不放弃优秀学术研究所依赖的判断力、严谨性与诚信。
核心原则
AI 可以自动化劳动，但不能替代学术责任。 请把每一份 AI 输出都当作草稿、清单、批评意见或编程辅助——在用于研究之前，对照原始来源核查论断、引用、系数、代码与公式。始终遵守所在机构、资助方、期刊与数据提供方的规则；当这些规则更严格时，遵守更严格的规则。
语言
完全双语——English 与 中文（简体）。
许可与引用
基于 MIT License 发布。如果本仓库对你的研究或教学有帮助，请使用附带的 CITATION.cff 引用，或引用 SSRN 工作论文。
联系方式
- 领英 LinkedIn: Chaojie (Jay) Liu
- 小红书: SuperJay
反馈与贡献
欢迎提出问题、修正或新的 skill 建议：jay.liu@bristol.ac.uk——邮件主题 [AI Econ Finance Handbook] Question or correction。
Released by Chaojie (Jay) Liu · v1.0.1
AI_EFR_v1.0.0
AI for Economics and Finance Research — v1.0.0
The first public release of a practical, GitHub-native handbook and resource library for using AI and large language models responsibly and effectively in economics and finance research.
This release brings the repository to a complete, stable, and citable v1.0.0: a full bilingual handbook, a library of copy-ready research skills, agent and automation workflows, worked examples, source notes, and teaching materials — all built around one simple loop: read, copy a skill, use it on a real task, verify the output, and record what changed.
What's included
- 01 · Start Here — The Handbook. One consolidated reading guide covering the mental model, the maturity ladder, what AI is and isn't good at, the econ/finance research workflow, and data, GitHub, and project safety.
- 02 · Copy and Use — Skills and Templates. 13 ready-to-use skill files: research ideas, paper drafting and citations, economics and finance empirical methods, coding and debugging, literature review, causal inference and time series, theory and math, referee reports, and more.
- 03 · Set Up Agents and Workflows. 6 multi-step, Git-safe workflows for cleaning projects, the one-paper-one-repo setup, replication packages, update digests, parallel agents and worktrees, and GitHub review/publish.
- 04 · Examples, Diagrams, and Failure Cases. Worked asset-pricing and corporate-finance examples, a research-workflow diagram, and failure cases.
- 05 · Check Sources and Official Docs. A source-to-repo map, a source-use policy, and privacy-safe ways to stay updated.
- 06 · Teach, Practice, and Share Slides. Workshop outlines, a slide-ready talk structure, RA onboarding checklists, and presentation-practice activities.
Who it's for
Students, research assistants, PhD students, and faculty in economics and finance who want the genuine productivity gains of AI without surrendering the judgment, rigor, and integrity that good scholarship depends on.
Core principle
AI can automate labor, but not scholarly responsibility. Treat every AI output as a draft, checklist, critique, or coding aid — and verify claims, citations, coefficients, code, and equations against original sources before using them. Always follow your institution, funder, journal, and data-provider rules; where they are stricter, follow the stricter rule.
Languages
Fully bilingual — English and 中文 (Simplified Chinese).
License and citation
Released under the MIT License. If this repository supports your research or teaching, please cite it using the included CITATION.cff.
Connect
- LinkedIn: Chaojie (Jay) Liu
- 小红书 (RED): SuperJay
Feedback and contributions
Questions, corrections, or new-skill suggestions are welcome: jay.liu@bristol.ac.uk — subject line [AI Econ Finance Handbook] Question or correction.
#高效地将AI应用于经济与金融研究 — v1.0.0
本仓库的首个公开发布版本：一本实用的、以 GitHub 为载体的手册与资源库，帮助经济学与金融学研究者负责任且高效地使用 AI 与大语言模型。
此版本将仓库带到一个完整、稳定、可引用的 v1.0.0：包含完整的双语手册、可直接复制使用的研究 skills 库、agent 与自动化工作流、实例、来源说明，以及教学材料——全部围绕一个简单的循环展开：阅读、复制一个 skill、在真实研究任务上使用、核查输出、并记录改动。
包含内容
- 01 · 从这里开始 — 手册。 一份整合的阅读指南，涵盖思维模型、成熟度阶梯、AI 擅长与不擅长的事、经济/金融研究工作流，以及数据、GitHub 和项目安全。
- 02 · 复制与使用 — Skills 与模板。 13 个可直接使用的 skill 文件：研究想法、论文写作与引用、经济与金融实证方法、编程与调试、文献综述、因果推断与时间序列、理论模型与数学、审稿报告等。
- 03 · 搭建 Agents 与工作流。 6 个多步骤、Git 安全的工作流：清理项目、one-paper-one-repo 设置、复现包、更新摘要、并行 agents 与 worktrees，以及 GitHub review/publish。
- 04 · 实例、图示与失败案例。 资产定价与公司金融的实例、研究工作流图示，以及失败案例。
- 05 · 核查来源与官方文档。 来源到仓库的映射、来源使用政策，以及保护隐私的更新方式。
- 06 · 教学、练习与分享幻灯片。 工作坊大纲、可直接使用的演讲结构、RA 入门清单，以及演讲练习活动。
适合谁
经济学与金融学领域的学生、研究助理、博士生与教师——希望获得 AI 带来的真实效率提升，同时不放弃优秀学术研究所依赖的判断力、严谨性与诚信。
核心原则
AI 可以自动化劳动，但不能替代学术责任。请把每一份 AI 输出都当作草稿、清单、批评意见或编程辅助——在用于研究之前，对照原始来源核查论断、引用、系数、代码与公式。始终遵守所在机构、资助方、期刊与数据提供方的规则；当这些规则更严格时，遵守更严格的规则。
语言
完全双语——English 与 中文（简体）。
许可与引用
基于 MIT License 发布。如果本仓库对你的研究或教学有帮助，请使用附带的 CITATION.cff 进行引用。
联系方式
- 领英 LinkedIn: Chaojie (Jay) Liu
- 小红书: SuperJay
反馈与贡献
欢迎提出问题、修正或新的 skill 建议：jay.liu@bristol.ac.uk——邮件主题 [AI Econ Finance Handbook] Question or correction。
Released by Chaojie (Jay) Liu · v1.0.0