A tested Python toolkit demonstrating core techniques used in large-scale data and graph analytics: PCY frequent-pair mining, deterministic MinHash similarity, spectral graph partitioning, centrality, PageRank, and parameterized Neo4j Cypher queries.
- Two-pass PCY algorithm with hash-bucket candidate pruning
- Exact Jaccard similarity and a 128-permutation MinHash estimate
- Laplacian spectral bipartitioning through the Fiedler vector
- Betweenness centrality and PageRank on a flight network
- Safe Neo4j query examples with runtime parameters and no credentials
- Automated tests and deterministic output
py -m venv .venv
.\.venv\Scripts\python.exe -m pip install -r requirements.txt
$env:PYTHONPATH = "src"
.\.venv\Scripts\python.exe -m pytest
.\.venv\Scripts\python.exe examples\run_demo.pyThe demo writes its complete numerical report to outputs/analysis.json and
regenerates the visualization above.
| Path | Purpose |
|---|---|
src/bigdata_graph/ |
Reusable Python implementations |
examples/run_demo.py |
End-to-end reproducible example |
tests/ |
Automated correctness tests |
cypher/ |
Parameterized Neo4j graph queries |
outputs/ |
Generated report and visualization |
This repository consolidates practical work in big-data management and analytics into a clean, standalone project. The examples use small synthetic datasets so every result is reproducible locally; the implementations expose the same concepts used with distributed baskets and large property graphs.
Abdullah Msallam
MIT