Teacher guide — Grade 10, Unit 8: Project: Air in my city¶

Time: 2 lessons + project (about 4 periods). Track: B (project).

#	Lesson	Time
1	Plan and clean	40–45 min
2	Analyse and publish	40–45 min
P	Unit project: Air in my city	2 periods

Before you start¶

Materials: Notebook notebooks/air_in_my_city.ipynb.
Read docs/pedagogy/lesson-anatomy.md and docs/pedagogy/ai-use-policy.md.

If you have no computers / no internet¶

Printed monthly summary table; poster instead of dashboard.

Common misconceptions¶

One station represents the whole city.
Quoting health thresholds from memory; check the current official value and cite it.

Differentiation¶

Support: give a partially filled table / sentence starters; pair with a data-steward role.
Stretch: ask "what would change your conclusion?" and "who is missing?"

Answer key (check questions)¶

Lesson 1: Plan and clean 1. What must you cite for the limit?
Answer: The organisation and document it came from. 2. Why plan for missing data?
Answer: It affects counts and means. 3. What is a one-line question?
Answer: e.g. 'In which months is PM2.5 most often above the limit?'

Lesson 2: Analyse and publish 1. Three parts of a conclusion?
Answer: Claim, evidence, reasoning. 2. Name two limits in this study.
Answer: One station; missing days; (synthetic) data. 3. Why disclose AI use?
Answer: Honesty and fair assessment.

Assessment¶

Starter quiz each lesson (3 questions, one recalling the previous lesson); unit quiz in quiz.yml; project marked with the rubric in docs/pedagogy/assessment.md; 2-minute oral check per group.

Alignment¶

Optional Computer Science Grade 10: Python programming, databases and AI & contemporary technology; compulsory Maths data handling/statistics.

Answers: practice exercises¶

e.g. "How many days in Dec–Feb 2019–2020 had daily PM2.5 above Nepal's 24-hour standard in Kathmandu, compared with Jun–Aug?"
20 ÷ 25 = 80% of valid days. Dividing by 31 treats 6 unknown days as clean, understating the problem.
So changing a choice (e.g. a revised standard) updates every result consistently, and readers see the assumptions.
WHO 15 µg/m³: Jan–Apr and Nov–Dec 100%; Oct 98%; Sep 91%; May 83%; Aug 47%; Jul 44%; Jun 36% — below 100% in May–Oct.
2019: 209 of 348 valid days; 2020: 212 of 354 — about 60% both years; no meaningful difference (synthetic data has no lockdown effect built in).
The 7-day mean stays above 40 from about October to April (it crosses the line during April–May and September–October); it is below 40 throughout June–August.
Different handling of −999 (left in, or replaced by 0), different denominator (all days vs valid days), > vs >=, different file versions, different date parsing (day-first). Compare the logs line by line.
The data show when and how much PM2.5 was high, not why. Traffic counts, fuel use, brick-kiln and burning records, weather (inversions, wind), source-apportionment studies, or natural experiments (strikes, lockdowns) would help.

Real-data exercises: notes¶

R1 Real data show the same seasonal pattern; spring 2020 lockdown (late March–May) often shows lower PM2.5 than the same weeks in nearby years, though spring forest fires can mask it. Let students argue from their own numbers.
R2 OpenAQ station availability changes; Open-Meteo weather is always available.
R3 The GHO page gives modelled attributable deaths with uncertainty ranges — teach students to quote the range, the year and the source.