Sample-data demonstration

Which customers are about to leave

This shows our AI testing process on public sample data. You can't upload your own data here.

  1. 1

    Pick a sample dataset

    Public datasets, each a real business question.

  2. 2

    See how AI tested it

    The same process, already run on that data.

  3. 3

    Download predictions and results

    Every held-back row with its actual outcome and the prediction.

Step 2

See how AI tested it

Sample data
Iranian Churn (UCI)
Predicts
Churn
Task
Classification

Iranian Churn [Dataset] (2020). UCI Machine Learning Repository. https://doi.org/10.24432/C5JW3Z. CC BY 4.0. Modified: split into training and held-out rows, predictions added.

STRONG CERTIFIED — AUDITABLE & REPEATABLE PREDICTIVE MODEL

Its predictive value is STRONG, measured on held-out rows that took no part in choosing it.

Winner
Random Forest (best of 12 models built)
Trained on
2,835 rows

Certificate · holdout, 136 rows that took no part in choosing the winner

Balanced-accuracy lift
83.8% (90% range 63.6–100.0%)
Balanced accuracy
91.9% vs 50.0% for the baseline
Baseline
Always predict the most common training class (0)

Compare · all 272 held-out rows

The product marks these Compare figures as diagnostic: “These results do not establish predictive performance” (outcome timing was not declared for this run).

Correct
266 of 272

Run: model ed250472-ad72-4de5-bb95-ee506409faf1 · tournament iranian_churn_20261028-210459 · completed 2026-10-28T21:06:31Z

Step 3

Download predictions and results

Predictions (CSV)

272 held-out rows: the feature values, the actual outcome, the prediction and its confidence.

Download CSV
Results (text)

The verdict, the certificate and Compare figures, and where the data came from.

Download results

The rows are numbered 1 to 272 in the order the product tested them (heldout_row). These are not row numbers from the original dataset: the split shuffled the rows.

← Pick another sample dataset