Day 5 · Project finalization

Day 5 — Linguistic Data Analysis II

How to use this notebook

This is your single submission for the day. It has two parts:

  • Part A · Tutorial — assemble your full pipeline end-to-end and produce your final evaluation.
  • Part B · Corpus Lab — finalize your one-page report and notebook (to be written).

You only edit the cells marked ✏️ YOU EDIT. Cells marked 🔧 Library cell are pre-written — run them, don’t change them.

➡️ Work top to bottom. When you’re done, Runtime → Run all, then File → Download → Download .ipynb and submit that file.

Part A · Tutorial — assemble & finalize

By today you have all the pieces: a gold set (Day 4), a prompt you iterated (Day 3), and the evaluation pipeline (Day 2). Today you run them together, end-to-end, on your held-out gold set and capture the final numbers.

The four ready-made track notebooks in the lda2-final-template repo (notebooks/cefr.ipynb, raamove.ipynb, cars50.ipynb, l2_errors.ipynb) are the fully-worked versions of this pipeline — use the one matching your track as your starting point.

Warning🚧 TODO — to be written

The step-by-step finalization walkthrough is still being written. It will cover:

  • loading your own gold set from Drive,
  • running your best prompt end-to-end,
  • producing the final per-class P/R/F1 + confusion matrix,
  • and exporting the figures for your slides.

Part B · Corpus Lab — the one-page report

Your notebook is submitted alongside a one-page report. Draft its five sections here as you finish each analysis (full guidance: mini-project tracks).

Warning🚧 TODO — to be written

Report scaffold (to be expanded):

  1. Scheme & gold — label set; how you built the gold set; what QC changed.
  2. Prompt iterations — table of changes and F1 at each step.
  3. Evaluation — per-class precision/recall/F1 + confusion matrix.
  4. Error analysis — concrete misses; model’s fault or the scheme’s?
  5. Limitations — stochasticity, contamination risk, what the numbers don’t show.

✅ Before you submit

  1. Runtime → Run all and check every cell ran without error.
  2. Part A outputs are visible (tables / charts / the model’s answers).
  3. Part B self-check prints ✅ (or your TODO answers are filled in).
  4. File → Download → Download .ipynb and upload that one file.