What you claim
report/REPORT.md is the readable paper. Prediction and limitation are yours. The results block is generated.
A research report is not a number you type once. It is three files that stay in sync: Markdown you write, a Python script that computes, and a LaTeX document that typesets. Tonight you run that loop on your Week 2 or Week 3 CSV.
report/REPORT.md is the readable paper. Prediction and limitation are yours. The results block is generated.
report/scripts/summarize.py
reads the CSV and writes
output/numbers.md and
output/numbers.tex. Standard library only.
report/main.tex
\inputs the generated snippet. A PDF is a bonus if
pdflatex is present; the .tex is enough.
If your studio repo does not yet have a report/ folder,
copy it from this handbook
(report/README.md).
Put your CSV in data/. In a Codespace:
cd report python3 -m venv .venv source .venv/bin/activate bash build.sh ../data/YOURFILE.csv
Read the generated block in REPORT.md. Change one
sentence, or one line of the script (exclusion bounds, or a second
statistic). Run build.sh again. The optional
descriptives page is still here if you want
a histogram; it is not this week’s report.
Load a file, write a prediction, then rebuild.
Run the script to fill this block.
\input{output/numbers}
Arts: the prose. After the first build, rewrite the title, prediction, and limitation until a stranger can state the claim in one sentence. Do not invent a prettier mean.
Science: the script. Change the exclusion rule or
add a second descriptive (median, or first half versus second half).
Rebuild so output/ updates. Paste the old and new
generated sentences in the report.
Paths to report/REPORT.md (or
lab-notes/week04.md),
report/scripts/summarize.py,
report/output/numbers.md or
numbers.tex, and report/main.tex.
Name one change you made after the first build. Due Tuesday
29 September (no class on the 30th).
build.sh on a sample CSV, reads the PDF or Markdown, changes one line, runs it again.Report: the four paths, and one iteration you can describe. The mean you typed by hand does not count.