skills.rest
xjtulyc
xjtulycCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

time-series-analysis

Build end-to-end time series analysis workflows for forecasting and model comparison.

33|6|Updated Mar 17, 2026
Search Tags:#time series#garch#arima#var#stl decomposition#forecasting evaluation
One-click install
npx skills add https://github.com/xjtulyc/awesome-rosetta-skills --skill time-series-analysis-xjtulyc
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: time-series-analysis
Source: https://github.com/xjtulyc/awesome-rosetta-skills/tree/main/skills/03-mathematics/time-series-analysis
Command: npx skills add https://github.com/xjtulyc/awesome-rosetta-skills --skill time-series-analysis-xjtulyc

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill solves the problem of turning raw time series into reliable forecasts by identifying models, fitting them correctly, decomposing seasonality, and comparing predictive accuracy with rigorous metrics.

Core Features & Use Cases

  • Univariate ARIMA/SARIMA modeling: Use stationarity testing plus ACF/PACF-driven identification to fit ARIMA and generate forecasts.
  • Seasonality decomposition: Apply STL to separate trend and seasonal components for clearer interpretation and better modeling.
  • Multivariate dynamics with VAR: Model interactions across multiple correlated time series and analyze effects via Granger causality and impulse response functions.
  • Volatility modeling with GARCH: Capture time-varying variance (volatility clustering) with GARCH(1,1) and produce conditional volatility forecasts.
  • Forecast evaluation and comparison: Compute scale-free errors (e.g., MASE) and use Diebold-Mariano testing to compare competing forecasts.

Quick Start

Use the time-series-analysis skill to run stationarity checks, fit ARIMA/SARIMA, perform STL decomposition, and evaluate a 24-step forecast with MASE and (optionally) the Diebold-Mariano test.

Frequently Asked Questions about time-series-analysis

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

FAQPage Schema
How do I forecast time series data using ARIMA and evaluate the results?▼

To forecast time series data, you fit ARIMA or SARIMA models using statsmodels and pmdarima after running stationarity tests, then evaluate out-of-sample accuracy using MASE and Diebold-Mariano testing.

What is the best way to model time-varying variance and volatility clustering in financial time series?▼

The best way to model time-varying variance is using GARCH(1,1) estimation with the arch library, which captures volatility clustering and produces conditional volatility forecasts for your time series.

How do I decompose a time series to separate trend and seasonal components?▼

To decompose a time series, apply STL seasonal decomposition to separate and extract trend and seasonal components, resulting in clearer data interpretation and improved subsequent modeling accuracy.

Can I model interactions across multiple correlated time series simultaneously?▼

Yes, you can model interactions across multiple correlated time series simultaneously by estimating Vector Autoregression (VAR) to analyze Granger causality and impulse response functions.

What stationarity tests do I need to run before fitting ARIMA models?▼

Before fitting ARIMA models, you need to run ADF and KPSS stationarity tests and inspect ACF/PACF plots to correctly identify the autoregressive and moving average parameters for your time series.

How do I compare competing time series forecasts rigorously?▼

To compare competing forecasts rigorously, compute scale-free error metrics like MASE and apply Diebold-Mariano testing with appropriate loss functions to statistically evaluate out-of-sample predictive accuracy.

Related Skills

Explore More Data & Analytics Skills→
fall-development-robfall-development-rob
7

data-fred

Retrieve FRED macroeconomic time series and metadata for economic analysis.

Community
Intermediate
camoneartcamoneart
5

grafana-dashboards

Design and deploy Grafana dashboards for real-time system and application metrics.

Community
Intermediate
jasrajtulsijasrajtulsi

statsmodels

Fit statistical models and run diagnostics in Python with statsmodels.

Community
Advanced
brainbytes-devbrainbytes-dev
10

deep-learning-trading

Forecast financial time series price movements using LSTM, Transformer, or CNN-1D deep learning models.

Community
Advanced