Causal ML · the rare one
Live repoSynthetic dataRetail Growth Intelligence — Uplift Modelling
Independent research · synthetic data
View the repo →What this is
A genuine two-stage X-learner for campaign uplift with control-arm up-weighting and per-campaign propensity blending. Committed outputs, unlike most of the portfolio.
Metrics
Overall ATE
0.0444 mean response-rate uplift
Top decile
0.0754 (~1.7x average)
Qini-like area
744.8
vs 502 baseline
Churn ROC-AUC
0.844
vs LightGBM 0.812, paired bootstrap
Rank corr.
Spearman 0.406
Stack
Provenance
Synthetic data throughout. Figures demonstrate method, not a measured commercial outcome.
What I would not claim
Nothing beyond what is listed above.
Explore the uplift
Observed uplift per predicted decile and the Qini curve from the X-learner. Slide to target the top K% of customers and see the captured incremental response recompute.
+0.0444
Overall ATE
response-rate lift
+0.0754
Top decile
~1.7x the average
744.8
Qini area
vs random targeting
0.406
Spearman rank
predicted vs observed
Observed uplift by decile
Percentage-point lift in response rate, treatment vs control, within each predicted decile.
Qini curve
Cumulative incremental responders captured (pink) against random targeting (cyan).
31.1%
of total incremental response captured
236
incremental responders (model units)
1.55x
the random-targeting baseline
| Decile | Customers | Observed uplift (pp) | Cumulative captured responders |
|---|---|---|---|
| 1 | 2,052 | 7.54 | 126 |
| 2 | 2,052 | 6.36 | 236 |
| 3 | 2,052 | 4.83 | 318 |
| 4 | 2,052 | 1.47 | 344 |
| 5 | 2,052 | 2.30 | 384 |
| 6 | 2,051 | 5.66 | 481 |
| 7 | 2,052 | 5.32 | 574 |
| 8 | 2,052 | 5.40 | 667 |
| 9 | 2,052 | 3.03 | 717 |
| 10 | 2,052 | 2.64 | 760 |
Synthetic retail data; demonstrates method, not a measured commercial outcome. The Qini series is derived from committed per-decile counts; headline figures are read straight from the committed model-comparison output.
Data provenance. Synthetic retail data; demonstrates method, not a measured commercial outcome. Per-decile rows and headline figures come from committed retail-intelligence outputs (data/uplift-deciles.json).