SNMExplorer
A new way to discover clusters and patterns in your data — the algorithm finds natural groupings automatically, so you don't have to guess how many groups to look for. Drop a CSV (or use the demo data) and click ▶ Run Analysis to start.
How does it work?
SNM (Scientific Numerical Model) uses a Lie-algebra-based eigenanalysis on a graph Laplacian to detect cluster structure: z-score columns → Euclidean distance matrix → negative column-sums on the diagonal → eigendecomposition. The most-negative eigenvalues mark the strongest cluster patterns; eigenvector weights tell you which items belong to each cluster. Algorithm by Dr. Joseph E. Johnson, Distinguished Professor Emeritus, USC Physics.
Your Data
Upload a CSV and identify the columns
Drop CSV file here
or click to browse
Rows = items (patients, samples, foods…) · Columns = numeric measurements · 10+ rows recommended
Auto-populated after CSV loads
Display Filters
Pick a strictness preset, or drag the handles after running
📐 Clustering strictness
Different datasets benefit from different settings. Pick a preset and click Run — fine-tune below if needed.
−1000–0
Range auto-sets after Run · drag handles to filter visible clusters
0.05–1.00
Show items with |weight| in this range
Fine-tune individual filters override the preset above
★ pre-filters extreme rows before analysis
Off by default — can hide real clusters on small datasets
★ each item appears in only its strongest cluster