I design agentic AI simulations and network-analytic workflows that turn messy business data into causal structure and decision support.

• 🎓 M.A. Computational Social Science, University of Chicago (Aug 2025)
• 🔬 RA: LLM focus-group agent simulation; political-music bipartite network analysis
• 🧠 Specialties: Social Network Analysis (Visualization, ERGM, QAP) · LLM agent workflows · API-based data collection · reproducible analysis
• 🛠️ Stack: Python (pandas, NumPy, scikit-learn, matplotlib, networkx), R (ggplot, ERGM, QAP), SQL, Git/GitHub, Refinitiv Eikon API

• Semiconductor Supply-Chain Data Collector & Visualizations (pinned)
  API-driven pipeline that builds a buyer–supplier network (~160 firms, 500+ ties), with model-ready exports and visual analytics.
  → See the pinned repository below.

• M.A. Thesis: “When Structure Takes Over” (public)
  Network-level evidence (triadic closure) explaining persistent outsourcing post-shortage, beyond firm capacity/policy.
  → Thesis Link – Knowledge@UChicago

• Turning raw APIs → clean network datasets
• Building transparent pipelines with clear I/O contracts
• Applying ERGM/QAP to separate correlation from structure
• Delivering readable visuals that reflect the model

• LLM Agent Focus-Group Simulation (RA) — multi-agent conversation framework; prompt/role eval against human groups.
• Political-Music Bipartite Network (RA) — campaign songs × candidates; network visuals and predictive models.

• 🔗 LinkedIn: linkedin.com/in/wangcosmo280
• 📧 Email: Cosmo280@uchicago.edu