Run 32 time series forecasting models with pure SQL in DuckDB.
npx skills add https://github.com/DataZooDE/anofox-forecast --skill anofox-forecast
Please help me install this Agent Skill. Skill: anofox-forecast Source: https://github.com/DataZooDE/anofox-forecast/tree/main/.claude/skills/anofox-forecast Command: npx skills add https://github.com/DataZooDE/anofox-forecast --skill anofox-forecast
This Skill provides a comprehensive set of tools for time series forecasting directly within DuckDB, eliminating the need to move data to external Python environments for analysis.
Use the anofox-forecast skill to forecast 14 days ahead using the AutoETS model on your 'sales' table, grouping by 'product_id' with 'date' as the time column and 'revenue' as the value.
High-intent search queries and answers about installing and using this skill.
Time series forecasting in SQL is achievable directly within DuckDB using pure SQL queries. This approach eliminates the need to export data to external Python environments, supporting 32 forecasting models, feature extraction, and cross-validation natively.
Yes, you can forecast multiple time series simultaneously in DuckDB. The workflow supports multi-series forecasting, allowing you to group data by identifiers like 'product_id' to efficiently generate predictions for thousands of items in a single query.
SQL time series analysis supports 32 forecasting models, ranging from simple baselines to advanced methods. Available models include AutoARIMA, MSTL, TBATS, and AutoETS, providing options for various time series patterns and forecasting requirements.
DuckDB supports comprehensive time series validation through cross-validation and backtesting. The workflow includes data preparation, feature calculation, and evaluation metrics, ensuring robust model assessment and conformal prediction directly within the database environment.
The best way to forecast daily sales for thousands of products in SQL is by leveraging DuckDB's optimized analytical capabilities. You can run pure SQL queries to forecast ahead, grouping by product ID and using your date and revenue columns for accurate multi-series predictions.
You do not need Python to use ML models like AutoARIMA with your database data. Advanced forecasting models, including MSTL and TBATS, can be applied directly within DuckDB using pure SQL, keeping the entire machine learning workflow inside the database.
hung-phanGenerate classical time-series features like lags and rolling windows.
menkesuReframe tactical work into outcome-based strategic framing with planning horizons.
benjaminwesternQuery, transform, and analyze CSV, Parquet, and JSON files with SQL.
RamanEbrahimiAutomate time series analysis with Python algorithms for classification, forecasting, and anomaly detection.
