Summary#

Audit date: 2026-05-28
Lectures audited: 26
Categories audited: writing, math, code, figures, references, links, admonitions (JAX out of scope for this series)
Average overall score: 8.0 / 10
Average per-category scores: writing 7.2, math 7.3, code 8.5, figures 6.5, references — (none in scope), links 9.1, admonitions 8.7

Priority distribution#

Priority	Count	%
HIGH	3	11.5%
MEDIUM	0	0%
LOW	19	73.1%
NONE	4	15.4%

The three HIGH-priority lectures all triggered the rule “any in-scope category ≤ 4” via the Figures category. Their overall scores are 7.0–7.5 (within the LOW band on its own), but a single-category floor of 4 (matplotlib, pandas_panel, workspace) raises priority.

Top systemic issues across the series#

[qe-writing-006] — appears in 24 / 26 lectures and remains the dominant style violation across the series. Multiple files exhibit 10–18 instances each.
[qe-fig-005] (missing name: field on figure directives) — appears in ~20 / 26 lectures. Every {figure} / {image} directive observed lacks a name: field, preventing {numref} cross-referencing. This is the single largest gap in the Figures category.
[qe-fig-003] (embedded matplotlib titles via ax.set_title / plt.title / plt.suptitle) — appears in 5 lectures outside the exercise/solution exception: pandas_panel.md (×4), workspace.md (×2), matplotlib.md (×2 — but pedagogically motivated), jax_intro.md (×1). matplotlib.md and workspace.md use it in demo code; pandas_panel.md uses it in non-exercise body code (clear violation).
[qe-fig-001] (unnecessary figsize= overrides) — appears in 9 / 26 lectures: matplotlib.md (×4 — pedagogical), numpy_vs_numba_vs_jax.md (×1 — 3D plot), jax_intro.md (×1), pandas.md (×2 in solutions), pandas_panel.md (none — pandas_panel uses default size), python_oop.md (×1), writing_good_code.md (×2), numpy.md (×4 historical leftovers — actually re-checked, mostly in matplotlib code samples). Total occurrences ~15.
[qe-link-002] (direct URLs to sibling lecture series instead of {doc} cross-references) — appears in 3 lectures: about_py.md (×1, jax), numba.md (×1, intro), sympy.md (×3, intro + intermediate).

Top strengths#

[qe-code-002] (Unicode Greek in code) — universally adopted in math-heavy lectures (α, β, γ, δ, ϵ, μ, ν, ρ, σ, λ, θ, ξ, η, Δ all appear). Programming-series exemplar.
[qe-code-003] (non-Anaconda installs at top with hide-output) — 100 % compliant across 10 lectures with !pip install cells (autodiff, jax_intro, numba, need_for_speed, numpy, oop_intro, pandas, pandas_panel, numpy_vs_numba_vs_jax, scipy).
[qe-code-004] (qe.Timer for benchmarking) — used exemplarily across 6 lectures (numba 8+ blocks, numpy_vs_numba_vs_jax 17+ blocks, need_for_speed, jax_intro, scipy, numpy). No time.time() patterns and no tic/toc legacy code anywhere.
[qe-code-005] (no %timeit / %%timeit) — 100 % compliant. Series consistently uses qe.Timer instead.
[qe-code-006] (binary-package warnings) — N/A. No graphviz / similar binary packages used in this series.
[qe-admon-002] (:class: dropdown for solutions) — 100 % compliant across all 17 lectures with solutions. Every solution-start directive observed in the audit corpus carries :class: dropdown.
[qe-admon-005] (solutions cross-reference exercises via :label:) — 100 % compliant. Every solution-start <label> matches a corresponding exercise-start :label: <label>.
[qe-link-002] (intra-series {doc} cross-references) — used cleanly in ~13 lectures (autodiff, functions, jax_intro, numba, oop_intro, pandas_panel, python_advanced_features, python_by_example, python_oop, numpy, numpy_vs_numba_vs_jax, troubleshooting, writing_good_code).
References category — no {cite} used anywhere in the series, so qe-ref-001 is N/A across all 26 lectures. The series does not draw on bibliographic citations.

Lectures ranked by priority (lowest score first)#

#	Lecture	Writing	Math	Code	Fig	Ref	Link	Adm	Overall	Priority
1	matplotlib	7	N/A	7	4	N/A	9	8	7.0	HIGH
2	workspace	8	N/A	7	4	N/A	9	9	7.4	HIGH
3	pandas_panel	7.5	N/A	8	3	N/A	10	9	7.5	HIGH
4	about_py	7.5	N/A	8	6	N/A	7	9	7.5	LOW
5	names	6.5	N/A	9	6	N/A	9	N/A	7.6	LOW
6	sympy	7	6	8	N/A	N/A	7	8	7.2	LOW
7	getting_started	6.5	N/A	8	7	N/A	9	8	7.7	LOW
8	numpy	6.5	7.5	8	7	N/A	9	8	7.7	LOW
9	numba	7	8	9	8	N/A	7	8	7.8	LOW
10	pandas	7	N/A	8	7	N/A	8	9	7.8	LOW
11	writing_good_code	7	7.5	9	5	N/A	10	9	7.9	LOW
12	python_oop	7	6	9	7	N/A	10	9	8.0	LOW
13	scipy	6.5	7.5	9	8	N/A	9	8	8.0	LOW
14	debugging	6.5	N/A	8	9	N/A	9	8	8.1	LOW
15	functions	6.5	N/A	8	N/A	N/A	9	9	8.1	LOW
16	need_for_speed	7.5	N/A	9	6	N/A	9	9	8.1	LOW
17	python_by_example	6.5	7	9	7	N/A	10	9	8.1	LOW
18	troubleshooting	7.5	N/A	N/A	8	N/A	9	N/A	8.2	LOW
19	numpy_vs_numba_vs_jax	7	8	9	7	N/A	10	9	8.3	LOW
20	jax_intro	7	N/A	9	7	N/A	10	9	8.4	LOW
21	python_essentials	6.5	N/A	9	N/A	N/A	9	9	8.4	LOW
22	oop_intro	7	N/A	8	N/A	N/A	10	9	8.5	LOW
23	autodiff	8	8	9	8	N/A	10	9	8.7	NONE
24	status	8.5	N/A	9	N/A	N/A	N/A	N/A	8.8	NONE
25	python_advanced_features	8	N/A	9	N/A	N/A	10	9	9.0	NONE
26	intro	9	N/A	N/A	N/A	N/A	10	N/A	9.5	NONE

Series-level recommendations#

Heading sweep remains the highest-leverage action. a global pass converting all H2/H3/H4 section headings from Title Case to sentence case (qe-writing-006). This alone would lift most of the LOW-priority lectures into NONE.
Figure metadata pass. Add name: fields (e.g., fig-jp-demo, fig-htop-parallel-npmat, fig-pandas-vs-rest) to the ~50+ {figure} and {image} directives across the series so they become {numref}-able (qe-fig-005). This is mechanical and high-impact.
Remove non-pedagogical plt.title / ax.set_title. In pandas_panel.md four such calls appear in main-body code cells and should be replaced with figure-directive captions (qe-fig-003). Workspace.md needs its sine_wave.py demo cleaned up similarly. matplotlib.md and jax_intro.md are borderline (one is a Matplotlib tutorial, the other diagrams a PRNG tree) and may warrant an explicit “this is a demonstration, don’t do this in production” callout.
Cross-series link conversion. Replace the ~5 bare quantecon.org URLs in about_py.md, numba.md, and sympy.md with {doc} cross-references using the appropriate intersphinx prefix (jax:, intro:, intermediate:) per qe-link-002.
Normalize blackboard probability notation. Carried over: \mathbb E → \mathbb{E}, \mathbb P → \mathbb{P} in numba.md, numpy.md, scipy.md, writing_good_code.md (qe-math-010, proposed). Also: python_oop.md should switch \mathbf{1} → \mathbb{1}; python_by_example.md should replace \pi * r^2 with \pi r^2 (qe-math-012, proposed); sympy.md narrative should switch E(X) to \mathbb{E}[X] (the SymPy code E(X) may remain since it is a SymPy function name).
Gated-exercise consistency. Several lectures mix bare {exercise} with {exercise-start} / {exercise-end}. Where the body contains only prose/math, bare {exercise} is technically allowed by qe-admon-001, but uniform gated syntax improves authoring consistency. Affected: numba.md, numpy.md, python_essentials.md, scipy.md, sympy.md.
Preserve the strong Code-category baseline. The series already does extremely well on qe-code-002 (Greek unicode), qe-code-003 (top-of-lecture installs with hide-output), qe-code-004 (qe.Timer), and qe-code-005 (no %timeit). Future contributions should keep this discipline.