Move from a messy dataset to a defensible, stakeholder-ready notebook with one connected AI workflow.

AI Analyst Workflow is a five-skill agentic system for planning, building, communicating, and reviewing trustworthy data analysis. Each skill installs independently, but together they support the complete notebook lifecycle:

Analyst Planning — turn a dataset and goal into a testable thesis, decision question, evidence plan, metrics, and risk map.

Analyst Structure — design a coherent notebook that runs from top to bottom without hidden state or analytical sprawl.

Analyst Chunk — work safely with large tabular files through bounded-memory processing, early aggregation, and visible validation.

Analyst Memo — convert notebook evidence into a concise, stakeholder-ready analytical artifact.

Analyst Review — audit validity, execution, reproducibility, visual evidence, and publish readiness.

These skills grew out of the analytical workflow I used to complete my Green Energy Capstone and build NYC Open Data projects that attracted attention from NYCHA and other city organizations.

Students and researchers building evidence-based notebooks

Data and product analysts who want clearer, more defensible work

Civic technologists working with large public datasets

Designers and strategists translating analysis into decisions

AI practitioners who want reusable capabilities instead of one-off prompts

Five installable skill ZIPs, a quick-start guide, a recommended workflow, and a single-user commercial-use license.

The skills are designed for supported OpenAI skill environments, including ChatGPT and Codex. Jupyter, Python, pandas, and other analysis dependencies are not included. Users remain responsible for reviewing outputs and validating analytical claims before consequential use.