spec: add timeseries-forecast-uncertainty specification

[github-actions]

New Specification: timeseries-forecast-uncertainty

Related to #3188

---

specification.md

timeseries-forecast-uncertainty: Time Series Forecast with Uncertainty Band

Description

A time series plot that displays historical observed data followed by a forecast projection with confidence intervals or uncertainty bands. The plot clearly distinguishes between the historical period and the forecast period using a vertical line marker, with shaded bands representing different confidence levels (typically 80% and 95%). This visualization is essential for communicating prediction uncertainty in forecasting applications, helping stakeholders understand both the expected values and the range of possible outcomes.

Applications

• Visualizing sales and demand forecasting with confidence intervals for inventory planning
• Displaying stock price or financial indicator predictions with uncertainty ranges for risk assessment
• Communicating weather forecast uncertainty with multiple confidence bands
• Presenting economic projections with confidence intervals for policy decisions
• Illustrating capacity planning forecasts with uncertainty bounds for resource allocation

Data

• date (datetime) - Time index covering both historical and forecast periods
• actual (numeric) - Historical observed values (null/NaN for forecast period)
• forecast (numeric) - Predicted values (typically starts where actual ends, may overlap slightly)
• lower_80 (numeric) - Lower bound of 80% confidence interval
• upper_80 (numeric) - Upper bound of 80% confidence interval
• lower_95 (numeric) - Lower bound of 95% confidence interval
• upper_95 (numeric) - Upper bound of 95% confidence interval
• Size: 50-200 total points (30-150 historical + 10-50 forecast)
• Example: Monthly sales data with 3 years history and 6-month forecast from Prophet or ARIMA model

Notes

• Use a solid line for historical data and a distinct style (dashed or different color) for forecast
• Mark the forecast start with a vertical line or shaded region
• Use nested shaded bands for confidence intervals: darker for 80%, lighter for 95%
• Semi-transparent fills (alpha 0.2-0.4) allow bands to overlap without obscuring data
• Include a clear legend identifying historical data, forecast, and confidence levels
• Consider using a consistent color family (e.g., blue for historical, orange for forecast)

---

Next: Add approved label to the issue to merge this PR.

---

🤖 spec-create workflow