The Analyst and AI
Replace the hype framing with the one that explains both the wins and the failures. Install the production-vs-judgment stance, the thinking-partner discipline, the working model of AI capability, and the verification habit that the rest of the course builds on. Ships eight working assets including the Production-and-Judgment Frame, the Working Model for Analysts, the Analyst Task Map, and the Verification Calibration.
Understand why AI is a genuine multiplier for analysis specifically, across the whole arc from question to recommendation
Use AI as a genuine thinking partner for analytical work, not an answer machine — to sharpen your analysis rather than replace your judgment
Build an accurate working model of what AI is actually good and bad at, so you know when to lean on it and when to be skeptical
Turn the working model into an operational habit: calibrate how much to verify by the stakes and the verifiability of each use
Framing the Question
The analytical judgment that determines whether you answer the right question. Structuring ambiguous problems, scoping the analysis, and the framing discipline that keeps AI from optimizing for the wrong objective.
Translate a vague or imprecise stakeholder ask into the precise, analyzable question actually worth answering
Anchor every analytical question to the decision it is meant to inform, so the analysis is actionable rather than merely interesting
Turn the anchored question into a tractable analysis plan — the hypotheses, the data, the method, the approach
Recognize framing as the highest-leverage judgment in analysis — the wrong question answered perfectly is worse than worthless
Working with Data
Query, prepare, and know your data with AI while keeping verification and interpretation yours. Ships the Querying Practice, the Data-Preparation Practice, the Data-Knowing Practice, and the False-Trustworthiness Guard.
Take the full production leverage AI offers for query-writing — write, translate, debug, and optimize queries with AI — because this is the largest single production gain in the analyst's workflow
Take the mechanical leverage AI offers for data preparation — transformation, reshaping, type conversion — while recognizing that cleaning decisions are judgment calls the analyst makes, not the tool
Take the exploration leverage AI offers — generating distributions, statistics, cross-tabs, visualizations — to survey the dataset faster and more thoroughly than manual EDA
Recognize the false-trustworthiness risk: AI makes data work faster and cleaner-looking, which can make unsound data easier to produce and harder to notice
Analysis & Modeling
Choose valid methods, draw sound inferences, and make the causal judgment with AI — while keeping the method validity, statistical rigor, and causal reasoning yours. Ships the Method-Selection Practice, the Sound-Inference Disciplines, the Causal-Judgment Discipline, and the Defensibility Standard.
Use AI to suggest, explain, and execute analytical methods while the analyst owns whether the method is right for the question and the data
Hold the statistical rigor that separates a sound inference from a plausible-looking but unsound one
Hold the causal judgment — whether a relationship is causal or merely correlational — which is the hardest and most consequential inference in analysis
Establish soundness — across method, inference, and causation — as the standard the analyst holds the analysis to and can defend
Visualization & Communicating Findings
The communication discipline for analytical work: AI-assisted visualization and narrative that keeps the interpretive judgment yours. The verification habits for findings that go out as authoritative.
Use AI to generate charts and visualizations fast while the analyst owns whether the visualization represents the data honestly
Turn findings into an analytical narrative a decision-maker can grasp and act on, using AI to draft while the analyst owns the story
Communicate the uncertainty and limitations of an analysis honestly — the part of communication that is hardest and most often skipped
Establish honest communication — across visualization, narrative, and uncertainty — as inseparable from being trusted
From Analysis to Recommendation & the Analyst's Judgment
The flagship module: assembling the full judgment practice into the analyst's operating standard. From analysis to recommendation, the complete discipline for an analyst who is right and trusted.
Bridge from what the analysis found to what to do about it, the move that makes analysis actionable
Recognize the domain and business judgment that a sound recommendation requires beyond the data — the context that is not in the numbers
Hold to what the analysis shows even when it is unwelcome — the independence that is central to the analyst's value
Bring together the throughline of the whole course: across every part of analysis, AI carries the production and the analyst supplies the judgment