Does this skill require a TimeGPT API key?

A NIXTLA_TIMEGPT_API_KEY is only required if you choose to use TimeGPT for the initial forecast generation; it is not required for local StatsForecast models.

Can I use this with models other than TimeGPT?

Yes, it is fully compatible with StatsForecast models (like AutoETS) and other standard time series outputs as long as they follow the required CSV structure.

What confidence levels are supported?

The skill supports any confidence level between 0 and 1. Common settings include 0.90 for 90% confidence or 0.95 for 95% confidence intervals.

What is conformal prediction in the context of this skill?

Conformal prediction is a framework used by this skill to provide mathematically grounded coverage for prediction intervals, ensuring your forecast ranges are statistically reliable.

What files are produced after running the uncertainty script?

The skill generates a CSV containing the original data with upper and lower bounds, a PNG visualization of the confidence bands, and a JSON file with calibration metrics.

Nixtla Uncertainty Quantifier

Name: Nixtla Uncertainty Quantifier
Author: intent-solutions-io

byintent-solutions-io

0•

データサイエンスとML

Quantifies time series forecast uncertainty using conformal prediction to generate reliable prediction intervals and confidence bands.

The Nixtla Uncertainty Quantifier is a specialized skill for Claude Code that adds a critical layer of risk assessment to time series forecasting. By leveraging advanced conformal prediction techniques—including quantile and jackknife+ methods—it transforms standard point forecasts from models like TimeGPT or StatsForecast into actionable ranges. This tool is essential for data scientists and analysts who need to understand the statistical 'worst-case' and 'best-case' scenarios, providing the mathematical confidence intervals and calibration metrics required for robust financial planning, supply chain optimization, and demand modeling.

主な機能

01Automated calibration metrics including coverage and interval width analysis

02Visual representation of forecast uncertainty with confidence bands

03Conformal prediction interval generation using Quantile and Jackknife+ methods

04Seamless integration with TimeGPT and StatsForecast pipelines

05Comprehensive CSV outputs with lower and upper prediction bounds

060 GitHub stars

ユースケース

01Risk-aware financial budgeting and revenue forecasting

02Determining safety stock levels for inventory management under uncertainty

03Energy demand scenario planning using 95% confidence intervals

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add intent-solutions-io/plugins-nixtla nixtla-uncertainty-quantifier

For use in Claude.ai and ChatGPT

主な機能

01Automated calibration metrics including coverage and interval width analysis

02Visual representation of forecast uncertainty with confidence bands

03Conformal prediction interval generation using Quantile and Jackknife+ methods

04Seamless integration with TimeGPT and StatsForecast pipelines

05Comprehensive CSV outputs with lower and upper prediction bounds

060 GitHub stars

ユースケース

01Risk-aware financial budgeting and revenue forecasting

02Determining safety stock levels for inventory management under uncertainty

03Energy demand scenario planning using 95% confidence intervals

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add intent-solutions-io/plugins-nixtla nixtla-uncertainty-quantifier

For use in Claude.ai and ChatGPT