Can I incorporate external variables like marketing spend?

Yes, the skill supports multivariate forecasting and custom regressors, allowing you to factor in external drivers like marketing budgets, weather, or holidays.

Is it integrated with SpecWeave's documentation features?

Yes, all forecasting experiments, stationarity tests, and evaluation metrics are automatically synced into your repository's living documentation via SpecWeave increments.

How does it prevent data leakage in time series?

The skill strictly enforces temporal train/test splits instead of random splits, ensuring that the model never 'sees' the future during its training phase.

What forecasting methods does this skill support?

It supports a wide range of methods including statistical baselines like ARIMA, business-centric models like Prophet, and deep learning architectures like LSTM and GRU.

Does it provide uncertainty estimation?

Absolutely. It generates point forecasts alongside confidence intervals (typically 80% and 95%) to help users understand the range of probable outcomes.

Time Series Forecaster

Name: Time Series Forecaster
Author: anton-abyzov

byanton-abyzov

•

Data Science & ML

Builds sophisticated time-dependent predictive models using statistical methods, machine learning, and deep learning within the SpecWeave framework.

The Time Series Forecaster skill empowers Claude to design and implement advanced forecasting pipelines specifically tailored for temporal data. It moves beyond standard machine learning assumptions by enforcing rigorous best practices such as temporal train/test splitting, stationarity checks, and seasonality decomposition. Whether you need classic ARIMA models for statistical rigor, business-focused Prophet forecasts for holiday awareness, or deep learning-based LSTM sequences for complex non-linear patterns, this skill provides the necessary structure to predict future trends and quantify uncertainty. Every model is integrated into the SpecWeave increment workflow, ensuring that your forecasting logic is backed by automated tests, stationarity reports, and living documentation.

Key Features

01Full integration with SpecWeave increments for traceable model development

02Automated seasonality detection and trend decomposition

03Temporal-safe cross-validation and specialized time-series metrics

04Multivariate forecasting for complex, interconnected datasets

05Multi-model support including ARIMA, Prophet, and LSTM architectures

0613 GitHub stars

Use Cases

01Predicting financial market trends and volatility using technical indicators

02Analyzing IoT sensor data for predictive maintenance and anomaly detection

03Forecasting retail sales and inventory demand with holiday-aware models

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add anton-abyzov/specweave time-series-forecaster

For use in Claude.ai and ChatGPT

Download Skill