TutorSyntax v26 | Teacher guidance | Original enrichment exercises
Suggested stage: KS3 extension, GCSE preparation, post-16 independent project prep.

Data quality:
Raw rows: 16. Repeated ID: A08 (one repeated row). Missing count: A09.
Unusual observation: 130 (A13), investigate rather than automatically exclude.
After counting record ID A08 once, there are 14 numerical observations.
Mean including 130: 63.29 visits. Mean excluding 130: 58.15 visits.
An observation with missing visits is omitted from means, not silently converted to zero.

Model evaluation:
Train on days 1-16; hold out days 17-24. Fit using training data only.
Compare regression MAE and training-mean baseline MAE on the same eight days.
The predictors do not include rain (a potential confounding/omitted factor).
Claim only what the held-out fictional sample supports.

Classifier:
Training rows T01-T12. Unseen test rows E01-E08. Distances standardise
temperature by 20 and moisture by 100. k=1/3/5; report TP TN FP FN plus
precision and recall, with zero-denominator handling. A test set of 8 is too small
for claims of real-world accuracy or fairness.

Assessment questions (correct answer letter, A/B/C):
1. B — Missing is not the same as a measured zero; investigate the source.
2. A — An unusual observation may be real or erroneous; record your decision and evidence.
3. B — A test set must stay separate from fitting and model selection for a fair final evaluation.
4. C — A model should be compared with a simple relevant reference using the same test examples.
5. A — Absolute error is |50 − 42| = 8 hires.
6. C — A false negative is an actual positive that the model misses.
7. A — One mistake can change a small-sample metric substantially; do not generalise too broadly.
8. B — Explain sources, methods, comparisons and limitations so someone can repeat your findings.

Project feedback rubric — 0/1/2 each (maximum 12; formative only)
1. Clear, answerable question and stated audience.
2. Documented data source, units, permissions and quality checks.
3. Reproducible method and honest train/test separation.
4. Results correctly labelled with an appropriate comparison.
5. Two realistic limitations and a safety/ethics consideration.
6. Reflection and concrete next step, including declaration of assistance.
0 = missing/incorrect, 1 = developing, 2 = clear and evidenced.
Never call these official examination or EPQ assessment marks.
