README.md

Sample Data Sets

1. SNAP (Stanford Large Network Dataset Collection)

SNAP provides a variety of real-world graph datasets, including social networks, web graphs, and collaboration networks.

• Website: SNAP Datasets
• Example Dataset: Facebook Social Network

Steps:

1. Visit the SNAP datasets page and choose a dataset, such as the Facebook Social Network.
2. Download the dataset in the edge list format (facebook_combined.txt.gz).
3. Extract the file and rename it to graph.edgelist.
4. Place the file in the data/sample_graph_data/ directory.

Example Command (Linux/MacOS):

wget http://snap.stanford.edu/data/facebook_combined.txt.gz
gunzip facebook_combined.txt.gz
mv facebook_combined.txt data/sample_graph_data/graph.edgelist

2. Open Graph Benchmark (OGB)

OGB provides large-scale graph datasets suitable for benchmarking graph neural networks and other graph algorithms.

• Website: OGB Datasets
• Example Dataset: ogbn-arxiv

Steps:

1. Install the OGB package using pip:

   pip install ogb

2. Use the OGB library to download and convert a dataset to edge list format.

Example Python Script:

from ogb.nodeproppred import NodePropPredDataset
import networkx as nx

# Load the dataset
dataset = NodePropPredDataset(name="ogbn-arxiv")
graph, _ = dataset[0]

# Convert to NetworkX graph
G = nx.Graph()
G.add_edges_from(graph['edge_index'].T)

# Save as edge list
nx.write_edgelist(G, 'data/sample_graph_data/graph.edgelist')

3. NetworkX Built-In Datasets

NetworkX includes some simple built-in datasets that can be used for testing purposes.

Example Code:

import networkx as nx

# Load a sample graph
G = nx.karate_club_graph()

# Save as edge list
nx.write_edgelist(G, 'data/sample_graph_data/graph.edgelist')

4. Custom Synthetic Data

If you want to generate a synthetic graph for testing:

Example Code:

import networkx as nx

# Generate a random graph
G = nx.erdos_renyi_graph(n=100, p=0.05)

# Save as edge list
nx.write_edgelist(G, 'data/sample_graph_data/graph.edgelist')

Summary:

• Use SNAP or OGB datasets for real-world data.
• Use NetworkX’s built-in graphs or generate synthetic graphs for quick tests.
• Ensure that the file is saved in the data/sample_graph_data/ directory with the name graph.edgelist.