Automated Forecasting Engine

1. Configure Your Forecast

Target Data*

CSV with Date column and the target metric (e.g., TotalRevenue).

Target Column (optional)

If empty, the first column in the target CSV (after Date) is used.

Feature Data (Optional)

CSV with Date and external features (e.g., Ad Cost, Weather).

Event Calendar (Optional)

CSV with Date and EventName (e.g., Black Friday).

Forecast Horizon: 90 days

2. View Results

Your forecast results will appear here after running the engine.

2. View Results

Best Model

—

Chosen by cross-validation using the primary metric.

MASE (CV)

—

Lower is better. < 1 beats a simple last-week forecast.

RMSE (CV)

—

Emphasizes big misses. Same units as your revenue.

MAE (CV)

—

Typical average miss in business units.

Coverage (Val)

—

Share of validation points inside the 95% band.

Bias (CV)

—

0 is ideal. Negative = under-forecast.

WAPE (CV)

—

Portfolio-level % error for the whole period.

RAE (CV)

—

Relative to naive. < 1 is better.

95% CI

—

CI Width Q

—

Half-width used for the 95% band.

Legacy Metrics (sMAPE / MAPE) Show

sMAPE (CV)

—

Legacy percentage error; more stable near zero.

MAPE (CV)

—

Legacy % error; can mislead with very small actuals.

Ensemble Details Show

Type: —

Stacker Coefficients:

Model	Coefficient

—

Horizon Breakdown (Tail Holdout) Show

Horizon	MAE	WAPE

Training: — • Features: — • Horizon: —

Note: Future values for external features use seasonal (last year) fallback or last value. Event lead/lag effects are included. Confidence band: conformal (95%).

Download Chart

Sign in

1. Configure Your Forecast

2. View Results

2. View Results