1. Define the problem
Objective: translate a policy, research or business question into analytical scope. Outputs: research question, analytical scope and limitations.
How I Work
From question to reproducible evidence.
Problem framing, data audit, modeling, validation, visualization and documentation.
This workflow explains the public portfolio's operating logic. Power BI and Tableau remain planned/build-package-ready where no real artifact has been published.
Workflow
Ten-stage analytical process.
2. Understand stakeholders
Objective: identify users and review paths. Outputs: audience map, review path and communication format.
3. Audit data
Objective: check public status, privacy boundaries and quality. Outputs: data inventory, privacy status and quality warnings.
4. Clean and harmonize
Objective: prepare consistent analytical structures. Outputs: clean panel, variable catalog and artifact catalog.
5. Model with SQL/DuckDB
Objective: document analytics-ready schemas and validation queries. Outputs: DDL, marts and validation SQL.
6. Analyze with Python/R/Stata
Objective: use fit-for-purpose tools for indicators, models and reports. Outputs: indicators, models, tables and reports.
7. Validate
Objective: make data-quality and publication-status checks visible. Outputs: validation reports, data-quality reports and known warnings.
8. Visualize
Objective: create dashboards, explorers or reports matched to audience. Tools: HTML/JS, Python, R; Power BI and Tableau remain planned/build-package-ready.
9. Communicate
Objective: translate analysis into recruiter, research, policy or executive narratives. Outputs: case studies, policy briefs, papers and public reports.
10. Document and reproduce
Objective: leave a public trail for methods, citations and reuse. Outputs: README, methodology, citation metadata and release notes.