A basic graph traversal script using the Github API. Starting from a specified root-node Github user, this will traverse everyone they are following. This repeats in a breadth-first pattern until a threshold is reached.

github_traversal.py

import requests
import getpass
import sys
import json
import Queue

# This is a script, let's be lazy. We'll fill up this global and print it.
g = {"nodes": {}, "edges": []}
# And here's the cutoff criterion
MAX_NODES = 1000

# Log in to avoid rate limiting
un = raw_input("Username: ")
pw = getpass.getpass()
r = requests.get("https://api.github.com", auth=(un, pw))
if r.status_code != 200:
    sys.stderr.write("Could not authenticate with Github. Exiting.\n")
    sys.exit()
sys.stderr.write("Connected and authenticated.\n")

# BFS. Kick this ish off
queue = Queue.Queue()
queue.put((un, None))
while queue:
    if len(g["nodes"]) > MAX_NODES:
        break
    user, parent = queue.get()
    if user in g["nodes"]:
        continue
    sys.stderr.write("Traversing %s.\n" % user)
    in_deg = requests.get("https://api.github.com/users/" + user,
                          auth=(un, pw)).json()["followers"]
    g["nodes"][user] = {"followers": in_deg}
    if parent:
        g["edges"].append({"source": parent, "target": user})
    followed = [res["login"] for res in requests.get("https://api.github.com/"
                "users/%s/following" % user, auth=(un, pw)).json()]
    for f in followed:
        queue.put((f, user))

# Save this result to a file
with open("followed.json", "w") as out_file:
    out_file.write(json.dumps(g))