Commit `982249b041`

982249b041600983b0d233d6c93f74597ad49234

parent: a70cee9b1e

Unsigned

cmc <hello@cleberg.net> · 2026-08-07 02:13 UTC

Make sampling/sample.py reproducible and point to the real tool

Seed the example draw and note that audit_sample.py / sampling_tool records
population hash, seed, method, and version for defensible sampling.

Layout: unified · split

sampling/sample.py +12 −2

		@@ -1,5 +1,11 @@
1	1	"""
2	2	Creates a sample from a CSV or Excel file based on user-defined SAMPLE_SIZE.
	3
	4	NOTE: This is a minimal teaching snippet. For real fieldwork use the
	5	`audit_sample.py` CLI (or the `sampling_tool` package), which records the
	6	population hash, seed, method, and tool version in a manifest so the sample is
	7	reproducible and defensible. This file fixes a SEED only so the example itself is
	8	repeatable; it does not emit that provenance.
3	9	"""
4	10
5	11	# Import packages
		@@ -8,6 +14,10 @@ import pandas as pd
8	14	# Define the sample size
9	15	SAMPLE_SIZE = 25
10	16
	17	# A fixed seed makes the draw reproducible: same population + same seed => same
	18	# rows. Record the seed alongside any sample you rely on.
	19	SEED = 20260707
	20
11	21	# Import the data to a pandas DataFrame
12	22	df = pd.read_csv("FILENAME_GOES_HERE.csv")
13	23
		@@ -19,7 +29,7 @@ df = pd.read_csv("FILENAME_GOES_HERE.csv")
19	29	print("Dataframe size (rows, columns): ", df.shape)
20	30
21	31	# Sample
22		sample = df.sample(SAMPLE_SIZE)
	32	sample = df.sample(SAMPLE_SIZE, random_state=SEED)
23	33	print("Sample size: ", SAMPLE_SIZE)
24	34	print("Sample:\n", sample)
25	35
		@@ -31,4 +41,4 @@ print("Sample:\n", sample)
31	41	#
32	42	# # Sample Size: 25 + 5 replacement samples
33	43	# SAMPLE_SIZE = 30
34		# sample = df.sample(SAMPLE_SIZE, replace=True)
	44	# sample = df.sample(SAMPLE_SIZE, replace=True, random_state=SEED)

Commit 982249b041

Commit `982249b041`