THE RESULT
What the data shows
The largest computed metric is 2,231 for M; the smallest is 1 for 15056P. Metric: p90_to_p10 (Price ratio).
Each line has equal weight. This is a pricing-review screen, not a fraud detector.
| stockcode | lines | p10 | median | p90 | p90 to p10 |
|---|---|---|---|---|---|
| M | 321 | 0.25 | 1.65 | 557.7 | 2,230.88 |
| 37,340 | 136 | 0.39 | 0.69 | 8.29 | 21.26 |
| 37,370 | 166 | 1.06 | 1.25 | 16.13 | 15.22 |
| 16,219 | 100 | 0.06 | 0.83 | 0.83 | 13.83 |
| 22,945 | 177 | 0.85 | 0.85 | 10.88 | 12.8 |
| 85071C | 106 | 0.39 | 0.39 | 4.96 | 12.72 |
| 22,915 | 185 | 0.42 | 0.42 | 4.96 | 11.81 |
| 22,068 | 147 | 0.39 | 1.65 | 4.13 | 10.59 |
| 22,130 | 113 | 0.85 | 0.85 | 8.434 | 9.922 |
| 21,700 | 200 | 0.85 | 1.38 | 8.29 | 9.753 |
| 22,500 | 101 | 1.25 | 4.95 | 10.79 | 8.632 |
| 22,268 | 164 | 0.19 | 0.63 | 1.63 | 8.579 |
THE METHOD
From source to answer
Compare 10th and 90th percentile positive unit prices, with at least 100 lines per code.
The Python source exposes this study’s transformations. The complete project download includes shared preparation and evaluation routines.
THE NEXT DECISION
What follows from the finding
Review discounts, pack sizes and charges before labeling any price variation erroneous.
Where the conclusion stops
Historical invoice lines; credits are not reliably matched to original sales. Gross purchases are not profit. Unidentified customers cannot support customer-level conclusions. Exact repeated lines remain unless the study explicitly compares removal.
Related studies may reuse observations or holdouts. These are historical analyses; associations and backtests do not demonstrate commercial impact. Further model tuning needs new, untouched evaluation data.