Problem statement

Leadership asks three questions: Is it getting worse? Who’s affected? What’s the baseline? Answer with three charts from cloudnote.db, each tied to a hypothesis.

Outcomes

Choose chart types deliberately; label axes; state denominator and time window; name one weakness per chart.

Deliverables

• Notebook or script + 3 chart files (PNG/PDF) + interpretation.md (≤ 200 words total).
• Extend or fork explore_metrics.py.

Links

• Phase 3

• [ ] Matplotlib or Plotly quickstart (one ecosystem).
• [ ] One article on metric hygiene (cohorts, ratios, Simpson’s paradox) — skim.

Time: ~2 h.

Task A — Three plots (required)

From cloudnote.db, produce:

1. Trend or “getting worse” — e.g. event rate or tickets over time (state the window).
2. Segmentation — who is affected (segment, country, or platform).
3. Baseline / comparison — vs another segment or prior period.

Save charts as plot1.png, plot2.png, plot3.png (or PDF).

Task B — Interpretation

interpretation.md: for each chart, one sentence takeaway + one sentence weakness/limitation.

Task C — Optional LLM

Ask an LLM to draft matplotlib code; run and fix until correct. Note one bug the model introduced.

Submit: my-submissions/week-06/ with script or notebook.

30 min.

1

When do you use a line chart vs a bar chart for PM narratives?

2

What is wrong with a chart titled “Revenue” with no axis labels and no time range?

3

Define one falsifiable hypothesis about CloudNote using only concepts from dim_user / events / tickets—then name the evidence that would disprove it.

Self-grade: Phase 3 checklist.

Concepts & curated links: read the Learn chapter Visualization first (topic path, not calendar-only). This file is the weekly practice bank — extra drills and a checklist on top of Readings → Practice → Exam.

1. Practice bank

2. Week checklist

• [ ] Matplotlib/plotly skim
• [ ] P1–P2 + PRACTICE.md + EXAM.md