Module 03 Summary

What this module established

numpy and pandas disagree about the standard deviation by default — np.std uses ddof=0 and Series.std uses ddof=1. On 61 delivery times that is 26.74 against 26.97, and at n = 10 it is about 5%.

Carry forward

• On skewed measures the mean describes a case that does not exist. One value in 61 nearly halved the standard deviation and moved the median by 0.35.
• groupby drops rows whose KEY is NaN by default — 82 rows in, 81 counted. Pass dropna=False when a missing key is meaningful.
• Two groups with means of 34.75 and 35.21 had completely different shapes; one was bimodal. Plot before testing.
• Overlapping error bars do not mean no difference: the lesson's two intervals overlapped while the difference interval was [0.87, 9.13].

Before moving on

Move on when every sentence you have written names its measure, size, comparison and period.

Mark complete