Parameterized Parallel Graphlet Decomposition Library

Compute graph motifs fast

PGD is a parameterized library for parallel graphlet decomposition (also known as motif counting) with many flexible interchangeable components (e.g., ordering strategies, representation, approximate/exact variants, etc.).
It is fast, parallel, parameterized, modular, and easy-to-extend library for efficient graphlet counting.

About

Installation

First, clone a copy of the main git repo by running:

git clone https://github.com/nkahmed/pgd.git
cd pgd && make
./pgd -f ../data/socfb-Texas84.mtx 

The built version of pgd will be put in the main pgd/ directory

Tutorials

Nothing helps more than toying with pgd code and the flexible components directly, here's a basic tutorial to get you started in seconds.

Read more »

Documentation, API, Analysis

PGD is easy to use and extend --- supports multiple graph formats, and has many other features that make it easy to use within seconds with little effort.

Go to documentation »

Graphlet Sensemaking and Exploratory Analysis

Interactive graphlet decomposition for sensemaking and spotting large stars quickly Interactive graphlet decomposition for sensemaking and spotting large stars quickly Interactive graphlet decomposition for sensemaking and spotting large stars quickly

Additional Examples & Use Cases