Data & Method
The file is clean, and its signal is real, but one key is broken
Zero missing cells, zero duplicates, zero type errors across 5,000 rows and 60 columns.
26 of 33 significance tests survive a Bonferroni threshold. The limit is structural: Asset_ID does not
identify one asset, so nothing here can be tracked per asset over time.
⚠
Stage 1 cleaning checklist study
Playbook §2. Nothing was dropped or imputed.
Do the columns mean what they say? study
Playbook §18 step 4.
Significance register study
One-way ANOVA confirmed by Kruskal-Wallis; both must clear the Bonferroni threshold. η² ≥ 0.01 small, ≥ 0.06 medium, ≥ 0.14 large.
Columns piled at a bound study
≥ 5% of rows at the minimum or maximum. Treat ceiling-heavy indices as capped, not as measured beyond the cap.