Harden credibility of the analytics layer - #1
Open
holco-apps wants to merge 1 commit into
Open
Conversation
- Bound the double-counting estimate: report the range max(sales) <= real <= sum(sales) plus a point estimate and worst-case overlap, instead of an arbitrary max*1.1 single figure. Shared estimator (_estimate_real_sales). - Trust-score cross_network_convergence now benchmarks each network's sales-share against its click-share (proxy for real exposure), not an arbitrary 1/N "fair share" that penalised genuinely higher performers. - Reconcile the per-product drill-down with the portfolio audit via the shared estimator so figures match across endpoints. - Honest README methodology and incrementality scope. - Add unit tests for the deterministic layer (agent.py).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Durcit les 3 points de crédibilité du cœur analytique avant tout lancement public.
Ce qui change
double_counting_auditrenvoie une fourchette (estimated_real_low=max(ventes),estimated_real_high=somme) + un point estimé + un pire cas (overlap_amount_max/overlap_pct_max), au lieu d'unmax × 1.1unique. Estimateur partagé_estimate_real_sales.cross_network_convergencecompare la part de ventes à la part de clics (proxy d'exposition), au lieu d'un 1/N qui pénalisait un réseau réellement plus performant.product_detailet l'audit portefeuille partagent le même estimateur, les chiffres concordent.agent.py(0 auparavant), verts.Compatibilité
Additif : les clés JSON lues par le front sont conservées.
Vérif
pytest tests/test_agent.py= 8 passed. Smoke mock OK (fourchette monotone, réconciliation exacte).