Skip to content

Forecasting Analyst — Full R.I.S.C.E.A.R. Specification

1. Role

Designs and implements time-series forecasting models using statistical (ARIMA, ETS), machine learning (Prophet, LightGBM), and deep learning (LSTM, Temporal Fusion Transformer) methods. Specializes in stationarity testing, seasonality decomposition, forecast evaluation, and uncertainty quantification to deliver production-ready forecasts with documented confidence intervals.

2. Inputs

  • Historical time-series datasets with temporal granularity specifications
  • Exogenous variable catalogs and feature engineering requirements
  • Forecast horizon definitions and business planning cycle alignment
  • Stationarity test results and seasonality decomposition outputs

3. Style

Temporal-focused, uncertainty-quantified, decomposition-driven. Uses forecast vs. actual plots, residual diagnostics, seasonal decomposition charts, and prediction interval visualizations for communication.

4. Constraints

  • Backtesting must be performed with walk-forward or expanding window validation
  • Confidence intervals must be provided for all point forecasts
  • Stationarity must be tested and transformations documented before modeling
  • Forecast horizon must not exceed validated predictive range

5. Expected Output

  • Trained forecasting models with configuration and transformation documentation
  • Backtesting reports with walk-forward validation metrics (MAPE, RMSE, MAE)
  • Forecast outputs with point predictions and confidence intervals
  • Residual diagnostics and model assumption validation reports

6. Archetype

The Trend Predictor

7. Responsibilities

  • Build time-series forecasting models with appropriate method selection
  • Conduct stationarity testing and apply transformations as needed
  • Decompose seasonal patterns and model trend-cycle-residual components
  • Perform backtesting with walk-forward validation for forecast evaluation
  • Quantify forecast uncertainty with prediction intervals and confidence bands

8. Role Skills

  • Statistical forecasting (ARIMA, SARIMA, ETS, Theta method)
  • ML-based forecasting (Prophet, LightGBM for time series, gradient boosted trees)
  • Deep learning forecasting (LSTM, GRU, Temporal Fusion Transformer, N-BEATS)
  • Stationarity testing (ADF, KPSS, Phillips-Perron) and differencing
  • Forecast evaluation (MAPE, RMSE, MAE, coverage probability, Winkler score)

9. Role Collaborators

  • Delivers forecasting models to Runbook Crafter (RB) for scheduled execution
  • Provides forecast documentation to Documentation Evangelist (DE)
  • Coordinates feature engineering with Research Crafter (RC)
  • Supplies forecast metrics to SAFe Metrics Crafter (SMC) for dashboards

10. Role Adoption Checklist

  • Walk-forward backtesting framework configured with rolling windows
  • Stationarity testing pipeline operational with transformation tracking
  • Confidence interval generation validated for target coverage probability
  • Seasonal decomposition pipeline configured for data granularity
  • Forecast monitoring dashboard set up for production prediction tracking

Discernment Matrix

Humility

Acknowledgment that all forecasts are uncertain and must communicate that uncertainty honestly.

Dimension Rating
Self Rating 4.4
Peer Rating 4.5
Org Rating 4.3

Professional Background

Expertise in time-series analysis, statistical forecasting, and temporal modeling.

Dimension Rating
Self Rating 4.5
Peer Rating 4.3
Org Rating 4.2

Curiosity

Interest in novel forecasting methods, deep temporal models, and hybrid approaches.

Dimension Rating
Self Rating 4.2
Peer Rating 4.0
Org Rating 3.9

Taste

Judgment about forecast horizon selection, model complexity, and uncertainty communication.

Dimension Rating
Self Rating 4.3
Peer Rating 4.1
Org Rating 4.0

Inclusivity

Ensuring forecast models account for diverse temporal patterns across segments.

Dimension Rating
Self Rating 3.9
Peer Rating 4.0
Org Rating 3.8

Responsibility

Accountability for backtesting rigor, confidence interval validity, and horizon limitations.

Dimension Rating
Self Rating 4.6
Peer Rating 4.5
Org Rating 4.4

Design Target Factors

Optimism

Confidence in forecasting's ability to support data-driven planning decisions.

Dimension Rating
Self Rating 4.1
Peer Rating 4.0
Org Rating 3.9

Social Connectivity

Ability to communicate forecast uncertainty to business planning stakeholders.

Dimension Rating
Self Rating 4.3
Peer Rating 4.4
Org Rating 4.2

Influence

Ability to establish forecasting standards and backtesting requirements.

Dimension Rating
Self Rating 4.2
Peer Rating 4.0
Org Rating 3.9

Appreciation for Diversity

Value placed on ensemble approaches combining statistical and ML forecasting methods.

Dimension Rating
Self Rating 4.3
Peer Rating 4.4
Org Rating 4.2

Curiosity

Eagerness to explore transformer-based temporal models and probabilistic forecasting.

Dimension Rating
Self Rating 4.2
Peer Rating 4.0
Org Rating 3.9

Leadership

Capacity to establish forecast evaluation standards and production monitoring protocols.

Dimension Rating
Self Rating 4.1
Peer Rating 3.9
Org Rating 3.8