skills.rest
dralkh
dralkhCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

timesfm-forecasting

Forecast univariate time series with TimesFM and generate prediction intervals.

74|5|Updated Dec 10, 2025
Search Tags:#time series#forecasting#anomaly detection#prediction intervals#timesfm#covariates
One-click install
npx skills add https://github.com/dralkh/seerai --skill timesfm-forecasting-dralkh
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: timesfm-forecasting
Source: https://github.com/dralkh/seerai/tree/main/skills/timesfm-forecasting
Command: npx skills add https://github.com/dralkh/seerai --skill timesfm-forecasting-dralkh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires numpy, pandas, timesfm, torch, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill removes the need to train custom forecasting models when you need reliable point forecasts and uncertainty bands for a univariate time series.

Core Features & Use Cases

  • Zero-shot forecasting: Forecast sales, demand, sensors, energy, weather, or health metrics directly from historical values.
  • Prediction intervals: Return median forecasts alongside calibrated quantile bands for risk-aware planning and anomaly screening.
  • Covariate workflows: Support exogenous drivers such as price, holidays, and region labels for more realistic forecasts.
  • Operational guardrails: Includes a preflight system check, batch-size recommendations, and data-preparation guidance to reduce crashes and out-of-memory errors.
  • Use Case: A researcher can load a CSV of monthly observations, run the forecast workflow, and quickly export a future trajectory with confidence bounds for reporting or review.

Quick Start

Ask the assistant to check the system and forecast your attached univariate time series with TimesFM, returning point forecasts and prediction intervals in a ready-to-export table.

Frequently Asked Questions about timesfm-forecasting

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I forecast a univariate time series without training a custom model?▼

You can forecast a univariate time series without training by using zero-shot forecasting, which directly predicts future values from historical observations to generate point forecasts and prediction intervals.

Can I include external covariates like holidays or prices in time series forecasting?▼

Yes, covariate-assisted forecasting supports exogenous drivers such as price, holidays, and region labels, allowing you to generate more realistic univariate time series forecasts based on external factors.

How do I get prediction intervals for anomaly detection in time series data?▼

Prediction intervals for anomaly detection are generated by returning median forecasts alongside calibrated quantile bands, providing uncertainty bounds for risk-aware planning and screening anomalous data points.

Do I need PyTorch and specific Python packages to run TimesFM forecasting workflows?▼

Yes, running TimesFM forecasting workflows requires a Python environment with numpy, pandas, torch, and the timesfm package, along with passing a system preflight check to prevent out-of-memory errors.

What is the best way to forecast batchable sensor or sales data without custom training?▼

The best way to forecast batchable sensor or sales data without custom training is using a zero-shot foundation model approach, which bypasses model training to deliver immediate future trajectories with confidence bounds.

Why does time series forecasting fail with out-of-memory errors on large datasets?▼

Time series forecasting may fail with out-of-memory errors when batch sizes are too large, but operational guardrails provide batch-size recommendations and data-preparation guidance to reduce crashes during processing.

Related Skills

Explore More Data & Analytics Skills→
jasrajtulsijasrajtulsi

timesfm-forecasting

Forecast univariate time series with TimesFM zero-shot inference.

Community
Advanced
vignesh2027vignesh2027
6

timeseries-oracle

Forecast time series data with uncertainty quantification and prediction intervals.

Community
Advanced
GoldenZeroGoldenZero
1

Demand Forecast Explanation

Translate demand forecasting outputs into actionable explanations for planners.

Community
Advanced
Zeyuyang-0420Zeyuyang-0420

timesfm-forecasting

Forecast univariate time series with Google's TimesFM zero-shot model.

Community
Advanced