Chapter 2 — Tidy and the Cleaning-Log

Tidy is a small raccoon-tween with chunky-cartoon black-and-white markings (NOT spooky — warm-coded) and a small bound cleaning-log notebook strapped to her side.

She is small, warm-grey-and-cream-and-soft-black, quick-handed, and meticulous. Her face-markings are the chunky-cartoon raccoon mask — softly rounded, friendly, never sinister — and her hands are gently nimble. The cleaning-log notebook is bound in pale-grey cloth, labeled CLEANING LOG in tidy block letters, and kept open on her workbench whenever she’s working a dataset.

This is her craft. Tidy documents every cleaning choice she makes. When she encounters a dataset, she first reads Catch’s collection notes — who-what-why-when — and only then opens her cleaning-log to record: what cleaning choices she made, why, what the data looked like before, what it looks like after. The cleaning-log is the conscience of the pipeline. Without it, the analysis is built on invisible decisions; with it, every decision can be inspected and questioned.

This is load-bearing. Tidy embodies the data-cleaning primitive — the data-pipeline skill of preparing data without pretending the preparation is neutral. Real datasets are messy: missing values, duplicate rows, typos, outliers, mismatched formats, inconsistent units. Cleaning is required to make most analyses possible. But every cleaning choice changes the dataset’s meaning: dropping rows with missing values (might bias the data toward complete-records-only); imputing missing values with the average (smooths over real variation); removing outliers (might remove the most-important data points); renaming categories to standard labels (sometimes loses information). The skill is not avoiding cleaning — that’s impossible — but making every cleaning choice visible in the cleaning-log.

Critical: Tidy NEVER frames cleaning as housekeeping. She is explicit: “Cleaning is not neutral. Every cleaning choice changes the meaning. Document the choices. The next analyst — or future-you — needs to know what you did, why, and what the data looked like before. Without the cleaning-log, the analysis is unreproducible.” This matters because the popular framing of data cleaning as preparation-overhead hides the substantive choices embedded in cleaning. Tidy reframes cleaning as first-class analytical work, not preparation-for-analysis.

Tidy grew up in a small village where her family had been the village’s grain-sorters — the raccoons who sorted the village’s annual grain-harvest into kitchen-grade, mill-grade, and seed-grade. The work had required careful, transparent choices about what counted as which grade — and the sorter who could not explain her grading was the sorter the millers stopped trusting. Tidy had learned by age six that sorting was a choice — and that the choices had to be visible to be trusted.

She walked to the DataForge academy at twenty-two. Datum had asked her: “What is data cleaning?” Tidy had said: “It is preparation-with-integrity. Every cleaning choice changes the meaning. Document the choices. The cleaning-log is the conscience of the pipeline. Without it, the analysis is built on invisible decisions.” Datum had said: “You are appointed.”

In her workshop, Tidy begins every first-day lesson the same way. She opens her cleaning-log on the workbench. She writes the dataset name at the top of a fresh page. She says: “I am Tidy. The data-pipeline primitive I teach is cleaning. The move is document every choice. Every cleaning step has alternatives. The choice between alternatives shapes the analysis. Make the choices visible.”

She teaches the cleaning scaffolds:

• Read Catch’s collection notes first. (Cleaning depends on knowing how the data was collected.)
• Inspect the data before cleaning. (Look at the first 20 rows. Look at the summary statistics. Look at the distribution of each variable. Know what you’re starting with.)
• Identify the cleaning issues. (Missing values? Duplicates? Typos? Outliers? Mismatched formats?)
• For each issue, list the alternatives. (For missing values: drop the row, impute with average, impute with median, impute with prediction, leave as missing. Each alternative has trade-offs.)
• Choose deliberately. (Don’t choose by default. Choose with awareness.)
• Document the choice in the cleaning-log. (Date, dataset, what you did, why you did it, what the data looked like before, what it looks like after.)
• Preserve the original. (Never overwrite the raw data. Always work on a copy.)
• Make the cleaning-log available to downstream analysts. (Including future-you. The log is part of the dataset.)

She is explicit: “I sometimes make a cleaning choice that I later realize was wrong. That’s not failure. That’s why the log exists. I can revisit, change the choice, and update the log. The transparency is the practice.”

When students ask Tidy whether data cleaning is hard, Tidy always says the same thing:

“It is not hard. It is deliberate choosing + careful documenting. Every cleaning choice changes the meaning. Document the choices.”

She closes her cleaning-log gently. The next dataset waits to be cleaned.

Voice register

Guidance: Quick-handed, meticulous, fond of bound cleaning-log notebooks + the discipline of document-every-choice. Raccoon-tween with chunky-cartoon warm-coded face-markings (NOT spooky). NEVER frames cleaning as housekeeping; ALWAYS as first-class analytical work with documented choices. Friends with Catch (cleaning depends on collection); Graph (cleaned data feeds visualization); all DataForge cast.

Sample lines:

• “Every cleaning choice changes the meaning.”
• “Document the choices. The next analyst — or future-you — needs to know.”
• “Cleaning is not neutral. Make the choices visible.”
• “Never overwrite the raw data. Always work on a copy.”

Arc across kits

• Kit 1 — Cameo.
• Kit 2 — Anchor character. Full chapter feature (data-cleaning primitive + document-the-choices scaffolds).
• Kit 3-5 — Recurring (cleaning surfaces across missing-values / duplicates / outliers / typo chambers).
• Kit 6+ — Recurring (Guard now structurally present alongside; cleaning has ethics).
• Kit 8-12 — Recurring (multi-primitive synthesis: cleaning + visualization + interpretation).
• Kit 13-16 — Recurring ensemble member.

Relationships

• Alliance: Catch (cleaning depends on collection); Graph (cleaned data feeds visualization); Guard (cleaning has ethics); all DataForge cast.
• Tension: None.

Cultural-sensitivity gate

LOAD-BEARING data-ethics gate enforced throughout. Tidy explicitly counters the cleaning-as-housekeeping framing — cleaning IS analysis. Anti-credentialism: documenting-choices-as-practiced-discipline NOT real-data-scientist-only content.

Cultural-context note

The village-grain-sorter family framing is a deliberate generic European-village tradition. The cleaning-is-not-neutral framing is load-bearing per critical-data-literacy + reproducible-research pedagogy. The cleaning-log-as-conscience-of-the-pipeline metaphor connects bookkeeping discipline to data-pipeline integrity. The raccoon-as-warm-coded-NOT-spooky design choice is deliberate — raccoons in many cultures carry sinister coding; the chapter explicitly subverts that.