Model Management

Manage your trained models: list them, inspect feature importance, check metrics, and manage your ML license.

List all models

dataspoc-lens ml models

Shows every trained model in your bucket with its target column, training date, and primary metric:

Model                    Target    Trained       AUC
churned_activity         churned   2026-04-10    0.91
ltv_customers            ltv       2026-04-08    0.87
fraud_transactions       is_fraud  2026-03-22    0.95

Explain a model

dataspoc-lens ml explain --model <model>

Shows feature importance, evaluation metrics, and training details for a specific model:

Model: churned_activity
Target: churned
Algorithm: GradientBoosting
Trained: 2026-04-10 14:32:00
Rows: 45,231

Metrics:
  AUC:       0.91
  Accuracy:  0.88
  Precision: 0.84
  Recall:    0.79

Top features:
  1. days_since_last_login    (0.23)
  2. total_sessions_30d       (0.18)
  3. support_tickets_open     (0.14)
  4. plan_type                (0.11)
  5. months_as_customer       (0.09)

License status

dataspoc-lens ml status

Shows whether ML is activated, the license tier, and expiration date.

Activate ML

dataspoc-lens ml activate <key>

Activates DataSpoc ML with a license key. Contact ml@dataspoc.com to obtain a key.

Model artifacts in the bucket

Each trained model produces three files stored at bucket/ml/models/<model>/:

File	Description
model.pkl	Serialized trained model
features.json	Feature definitions, transformations, and column type mappings used during training
metrics.json	Evaluation metrics (accuracy, AUC, RMSE, precision, recall, etc.)

These artifacts are read by ml predict to score new data and by ml explain to display model details.