Build a dataset and a baseline to beat

Collect 300 claim emails labelled lost, damaged or late

Reading your own data is the step people skip and the one that pays most. Note how many of your 30 were wrong.

Run this to check

You pass when

300 lines of text and label. Have a model write them if you have none, then read 30 yourself and fix wrong labels.

Files you write

• data/all.jsonl

Shuffle once with a fixed seed and split 240, 30, 30

Run this to check

You pass when

240, 30 and 30 lines. No email appears in two files.

Files you write

• data/train.jsonl
• data/valid.jsonl
• data/test.jsonl

Train word counts plus logistic regression and score it on the test rows

Run this to check

You pass when

An accuracy number and a 3 by 3 table of which label was mistaken for which.

Files you write

• baseline.py

Score a hosted model on the same 30 rows with no training

Run this to check

You pass when

A second accuracy number and its cost. Write both in RESULTS.md: this is the bar.

Files you write

• baseline.py
• RESULTS.md