Everything from Weeks 1–10, in twelve lines

Predict which hand calculation Python will disagree with

• A poll: "which of this semester's results will Python get differently?" (None — but std() will surprise you: pandas divides by n − 1, NumPy by n. The Week 2 trap, one last time.)
• Reveal: the machine does not know more statistics than you do. It only does the arithmetic faster, and it will happily compute the wrong test.

No install · lists, Series, DataFrames, read_csv

• Step-by-step with screenshots: open Colab, first cell, shift-enter. Why nothing needs installing and where the notebook is saved.
• The thirty prices as a list → a pandas.Series; the pizzeria table as a DataFrame; pd.read_csv on the course's own CSV.
• A card on the three imports that carry the whole week: pandas as pd, matplotlib.pyplot as plt, scipy.stats as st.

describe() · mean · median · std · quantile · hist · boxplot

• A code block per Week 2 section, each next to the widget it reproduces: s.describe() gives the five-number summary in one line.
• The n − 1 catch, made visible: s.std() (pandas, ddof=1) vs np.std(s) (ddof=0). Which one is the sample standard deviation, and why the default differs between the two libraries.
• plt.hist and plt.boxplot — the same pictures, in three lines.

ttest_1samp · ttest_ind · ttest_rel · chi2_contingency · f_oneway · OLS

• One block per test, each checked against the hand result from its week: the CHF 20 claim, lake vs station, before-and-after the menu change, toppings by neighbourhood, four branches.
• Confidence intervals: st.t.interval — and the same number the Week 6 widget drew.
• Regression with statsmodels.formula.api.ols: reading summary() line by line — coefficient, standard error, t, p, R².

Self-check quiz & where this goes next

• Eight questions: read four Python outputs and answer the exam-style question each one settles; plus the ddof trap and one "the code ran, the test was wrong" item.

Workbook fields

• w12_notebook — the Colab link (set to "anyone with the link can view") for the notebook that reproduces your project's numbers.
• w12_learned — one line on what the notebook told you that the hand calculation did not, and one thing Python made easy that would have been an hour by hand.

The second half of the session

When this week is written, each exercise below becomes a form on this page and a slide in the deck, generated from one array so the two can never disagree — and its solution opens when the instructor reveals it in class. Same engine as Weeks 1–2 (exercises.js).

• Exercise 1 · First cell  8 min
  open Colab, a list, a Series — and where it saved
• Exercise 2 · The DataFrame  12 min
  read_csv on the course data, .head(), .describe()
• Exercise 3 · Describe, reproduced  15 min
  mean, median, std, quantile — matched to your Week 2 answers
• Exercise 4 · The ddof trap  12 min
  pandas .std() vs numpy .std() — which one is the sample SD
• Exercise 5 · Two pictures  12 min
  plt.hist and plt.boxplot, three lines each
• Exercise 6 · Three tests  20 min
  ttest_1samp, ttest_ind, ttest_rel against your hand results
• Exercise 7 · chi2_contingency and f_oneway  15 min
  the Week 10 tables, reproduced
• Exercise 8 · Read summary()  18 min
  statsmodels OLS: coefficient, se, t, p, R² — line by line