What metrics does it use for model evaluation?

The skill provides comprehensive evaluation reports featuring SMAPE, MASE, and MAE, along with visual comparison plots for model performance analysis.

Which forecasting models are supported?

It supports a wide range of models including statistical baselines (AutoARIMA, AutoETS), machine learning models via MLForecast, and generative AI models via TimeGPT.

Can I use this for production demand planning?

Yes, the skill is designed to generate production-ready code, including cross-validation scripts and performance evaluation metrics suitable for enterprise demand planning.

Does this skill require a Nixtla API key?

A Nixtla API key is required to access TimeGPT foundation model features, but the skill can still generate local forecasting pipelines using StatsForecast and MLForecast without one.

What data format does the Nixtla TimeGPT Lab use?

The skill follows the standard Nixtla schema, requiring columns for a series identifier (unique_id), a timestamp (ds), and the target value (y).

Nixtla TimeGPT Lab

Name: Nixtla TimeGPT Lab
Author: intent-solutions-io

byintent-solutions-io

Data Science & ML

Generates high-performance time series forecasting pipelines and model benchmarks using the Nixtla ecosystem.

About

Nixtla TimeGPT Lab transforms Claude into a specialized time series forecasting assistant that leverages the full Nixtla ecosystem, including TimeGPT, StatsForecast, and MLForecast. It automates the development of end-to-end forecasting workflows—from data preparation and schema mapping to model selection, cross-validation, and performance evaluation. Whether you are performing M4 benchmarking, predicting sales demand, or deploying forecasting microservices via FastAPI, this skill ensures best practices and production-ready code generation based on state-of-the-art predictive modeling patterns.

Key Features

Comprehensive model benchmarking with metrics including SMAPE, MASE, and MAE
Standardized schema mapping and data preparation for time series identifiers
Direct integration with Nixtla's TimeGPT foundation model for zero-shot forecasting
0 GitHub stars
Automated forecasting pipeline generation using StatsForecast and MLForecast
Production-ready code generation for cross-validation and confidence interval estimation

Use Cases

Predicting future sales and inventory demand for retail or supply chain optimization
Benchmarking classical statistical models against modern machine learning foundation models
Building and deploying scalable forecasting microservices with FastAPI

About

Key Features

Comprehensive model benchmarking with metrics including SMAPE, MASE, and MAE
Standardized schema mapping and data preparation for time series identifiers
Direct integration with Nixtla's TimeGPT foundation model for zero-shot forecasting
0 GitHub stars
Automated forecasting pipeline generation using StatsForecast and MLForecast
Production-ready code generation for cross-validation and confidence interval estimation

Use Cases

Predicting future sales and inventory demand for retail or supply chain optimization
Benchmarking classical statistical models against modern machine learning foundation models
Building and deploying scalable forecasting microservices with FastAPI