Filter and analyze time-series sensor data with pywayne.dsp components.
npx skills add https://github.com/wangyendt/wayne-skills --skill pywayne-dsp
Please help me install this Agent Skill. Skill: pywayne-dsp Source: https://github.com/wangyendt/wayne-skills/tree/main/pywayne/dsp Command: npx skills add https://github.com/wangyendt/wayne-skills --skill pywayne-dsp
Digital signal processing is essential for turning noisy sensor data into usable information. This skill provides a cohesive toolkit for filtering, detrending, peak detection, and curve similarity to streamline preprocessing and analysis.
Load your sensor data and apply ButterworthFilter or OneEuroFilter to obtain a clean signal, then run peak_det and CurveSimilarity to compare patterns.
High-intent search queries and answers about installing and using this skill.
To reduce noise in raw sensor data, apply the ButterworthFilter or OneEuroFilter to smooth time-series signals. These methods process batch and real-time inputs to remove high-frequency fluctuations and reveal underlying patterns.
Dynamic time warping is used to measure curve similarity by comparing time-series sequences. The CurveSimilarity component aligns temporal patterns that may vary in speed, enabling accurate shape matching between different sensor signals.
Detect peaks and valleys in time-series sensor data using the peak_det function or FindSlidingWindowExtremum. These tools identify local maxima and minima within sliding windows to pinpoint specific signal events for batch processing.
Yes, you can use this signal processing toolkit for real-time monitoring. It includes the OneEuroFilter for adaptive smoothing and WelfordStd for online standard deviation computation, supporting continuous sensor data streams.
Available detrending methods for time-series data include linear, mean, and LOESS options via SignalDetrend. Applying these techniques removes baseline drift from sensor data to expose true underlying signal trends.
The best way to preprocess sensor data before pattern analysis is to sequentially apply ButterworthFilter for noise reduction, SignalDetrend for baseline correction, and then CurveSimilarity to compare the resulting clean signals.
AlterLab-IEUForecast univariate time-series data with zero-shot TimesFM point and quantile forecasts.
Fuuz-Industrial-IntelligenceBuild machine learning pipelines for industrial telemetry analysis on FUUZ.
GuitarAlchemistAnalyze and transform signals using FFT, filtering, wavelets, and Kalman filters.
NebulavenusGenerate deterministic HLSL shader noise utilities for procedural textures.
