How taha715/midi-era-clustering works
This repository packages a KNIME workflow for clustering classical MIDI compositions by era. The documented path is feature extraction from raw MIDI, CSV ingestion, filtering, normalization, PCA, and k-means, followed by visualization and evaluation against known labels. The tree contains node settings and workflow metadata, but no feature-extraction source or source excerpts; edges therefore reflect only README-documented flow, and the documented Python stage has no repository path.
Main entry points: workflow.knime KNIME workflow · PCA/settings.xml PCA · kMeans/settings.xml k-means · Scatter Plot/settings.xml Scatter Plot
Components 13
Input preparation
- Feature extraction
- CSV Reader
CSV Reader/settings.xml - Column Filter
Column Filter/settings.xml - Normalizer
Normalizer/settings.xml
Visualization and evaluation
- Scatter Plot
Scatter Plot/settings.xml - Color Manager
Color Manager/settings.xml - Joiner
Joiner/settings.xml - CSV Writer
CSV Writer/settings.xml
Other components
- MIDI dataset
- Era ground truth
- KNIME workflow
workflow.knime
Connections 17
- MIDI dataset → Feature extraction: extracts features
README.md - Feature extraction → CSV Reader: provides feature CSV
README.md - CSV Reader → Column Filter: ingests for
README.md - Column Filter → Normalizer: then filters
README.md - Normalizer → PCA: then normalizes
README.md - PCA → k-means: then reduces
README.md - k-means → Scatter Plot: visualizes clusters
README.md - k-means → Era ground truth: scores against
README.md - PCA → Scatter Plot: plots 2D projection
README.md - KNIME workflow → CSV Reader: orchestrates
README.md - KNIME workflow → Column Filter: orchestrates
README.md - KNIME workflow → Normalizer: orchestrates
README.md - KNIME workflow → PCA: orchestrates
README.md - KNIME workflow → k-means: orchestrates
README.md - KNIME workflow → Scatter Plot: orchestrates
README.md - Joiner → Era ground truth: may join labels
- Color Manager → Scatter Plot: may color clusters