Research toolkit
Resources
Learning Materials
Methods·Causal Inference
Causal Inference
Design-based and semiparametric tools for empirical research.
- fect / interflex / fdid (Xu) + Applied Empirical Methods (Goldsmith) — DiD/IV/RDD.
- Double Machine Learning (Ahrens et al., 2025) — Neyman-orthogonal moments + cross-fitting.
- Matching & Weighting (Imai) — propensity-score matching, IPW, covariate balance.
Methods·Computational
Computational Methods
Practical foundations for machine learning, language models, and agent simulations.
- The Missing Semester of Your CS Education (MIT) — shell, git, scripting, harness.
- Deep Learning for Economists — neural nets, sequence models, and LLMs.
- Natural Language Processing (CMU) — foundations of LLMs.
- Language Modeling from Scratch (Stanford CS336) — build and train LLMs from scratch.
- Agent-Based Simulation (Stanford CS222) — simulating social systems with LLM agents.
Economics·Empirical IO
Empirical IO
Industrial-organization tools and applications for marketing.
- Empirical IO (Conlon, NYU) — demand, mergers, productivity, dynamic choice.
- Handbook of the Economics of Marketing (Dubé & Rossi, 2019) — IO-for-marketing.
Economics·Information
Information Economics
Foundations for understanding information, search, and learning.
- Information Economics (Välimäki, 2014) — information economics basics.
- Digital Economics (Goldfarb & Tucker, 2019) — search, replication, tracking, verification costs.
- Information & Learning in Economic Theory (Liang, 2022) — the Blackwell order and learning.
Agent Skills
Reusable workflows·Research agents
Agent Skills
Structured workflows for common research tasks.
- Agent Configuration — configuring Claude Code for a research project.
- Web Scraping — structured web and social data collection.
- Literature Review — verified bibliographies via OpenAlex and Crossref.
- Big-Data Processing — Polars and DuckDB for data too large for Pandas.
- LLM Annotation — LLM data labels validated against a human gold-set.
- Visualization — publication figures in R + ggplot2.
- Slide Generation — Reveal.js reading-group decks from a paper.
- Conceptual Framework — source-grounded figures of a paper's constructs, relationships, and measurements.