# Research report pipeline

Week 4 laboratory. Keep this folder in the private studio repository.

The loop is: edit prose or the script → run the script → read the generated numbers → change something → run again. Do not type a mean by hand.

```
report/
  REPORT.md              # the markdown paper you write
  main.tex               # the same paper in LaTeX
  scripts/summarize.py   # reads a CSV, writes numbers
  output/numbers.md      # generated; do not edit
  output/numbers.tex     # generated; do not edit
  build.sh               # one command for the loop
```

## Run once

From this folder, in a Codespace terminal:

```bash
python3 -m venv .venv
source .venv/bin/activate
python scripts/summarize.py --csv ../data/YOURFILE.csv
bash build.sh ../data/YOURFILE.csv
```

`summarize.py` uses only the Python standard library. The virtual environment is so later packages (if you add any) stay off the system Python. Do not `pip install` unless the instructor says to.

If `pdflatex` is missing, that is fine tonight. The graded files are `REPORT.md`, `main.tex`, the script, and the generated snippets. A PDF is a bonus, not a requirement.

## The loop

1. Write a prediction in `REPORT.md` *before* you look at the generated mean.
2. Run the script.
3. Read `output/numbers.md`. If the n or the mean looks wrong, fix the CSV path or the exclusion rule in the script — do not edit `output/`.
4. Change one sentence of prose, or one line of the script.
5. Run again. Note what changed.

Copy `REPORT.md` to `lab-notes/week04.md` when you submit.