Informatics: The University of Melbourne

Science&Art: Social Networks

Social Networking

Application of Informatics

Data visualisation is a powerful tool to get insight from complex data, particularly from the vast collections of information available online. One example of this kind of data would be social networking data. In the last few years, we've all seen the explosion of social networking sites such as MySpace, Facebook, Hi5, LastfFM and many more. Many of them make available their 'social connectedness' data, giving interesting insights into how poeple relate.

Informatics in Action

In this example, we map the social connectedness of data from LastFM, an online social networking site which recommends music based on the preferences of like-minded people. Within complex relationship networks, visualization helps us to investigate the characteristics of social interactions: for example, whether a user has few friends who each have lots of friends, or he/she is someone with lots of friends, each of whom just have few friends.

Resources

Data

XML dataset of one social network from LastFM

Code

View the Python source code
Download the Python source code [Right click and Save As]

1import os
2os.environ[ 'TMP' ] = 'c:/sandbox/temp/'
3
4from mod_python import *
5import sys
6import urllib
7import xml.dom.minidom
8from array import *
9from xml.dom.minidom import Node
10from pylab import *
11from networkx import *
12from networkx.generators.atlas import *
13from minghi import *
14from ming import *
15
16
17nodeSize = {}
18textSize = {}
19lineSize = {}
20myCirc = {}
21
22proxies = {'http': 'http://wwwproxy.unimelb.edu.au:8000'} # unimelb proxy
23#proxies = {}  # use this if you want to by pass the proxy
24
25def findfriends(user):
26        # talk to lastfm web services
27        remotexml = urllib.urlopen( "http://ws.audioscrobbler.com/1.0/user/"+user+"/friends.xml",proxies=proxies )
28        friends = []
29        try:
30                doc = xml.dom.minidom.parse( remotexml )
31        except:
32                return friends
33        i = 0
34        for node in doc.getElementsByTagName("user"):
35                friends.append( node.getAttribute("username") )
36                i += 1
37                if i > 12:
38                        return friends
39
40        return friends
41
42def buildgraph(user, outputName ):
43        G = XGraph(name="lastfmgraph", multiedges=True, selfloops=False)
44
45        # build the graph
46        friendlist = findfriends( user )
47        nodeSize[user] = 80
48        textSize[user] = 240
49        lineSize[user] = 40
50        for u in friendlist:
51                G.add_edge( user, u )
52                friendlist2 = findfriends( u )
53                nodeSize[u] = 50
54                textSize[u] = 150
55                lineSize[u] = 20
56                for u2 in friendlist2:
57                        G.add_edge( u, u2 )
58                        if nodeSize.has_key(u2) == False:
59                                nodeSize[u2] = 20
60                                textSize[u2] = 60
61                                lineSize[u2] = 10
62
63        # layouting algorithm to generate the position for all nodes
64        pos = spring_layout(G,iterations=15)
65
66        # initialize flash file
67        Ming_setScale(1)
68        m = SWFMovie()
69        m.setDimension(6400, 4800)
70        m.setRate(12.0)
71
72        minx = 100
73        maxx = -100
74        miny = 100
75        maxy = -100
76
77        # preparing the scale for flash file
78        for v in G:
79                (px,py) = pos[v]
80                if px < minx:
81                        minx = px
82                if px > maxx:
83                        maxx = px
84                if py < miny:
85                        miny = py
86                if py > maxy:
87                        maxy = py
88        centerx = (minx+maxx)/2
89        centery = (miny+maxy)/2
90        lengthx = maxx-minx
91        lengthy = maxy-miny
92        xscale = 6400/lengthx
93        yscale = 4800/lengthy
94        if xscale > yscale:
95                scale = yscale
96        else:
97                scale = xscale
98        scale = 0.9*scale
99
100        newpos = {}
101
102        # draw nodes
103        for v in G:
104                (px,py) = pos[v]
105                px = (px-centerx)*scale + 3200
106                py = -(py-centery)*scale + 2400
107                newpos[v] = (px,py)
108                size = nodeSize[v]*3
109                myCirc = drawCircle( px, py, size, 200, 0, 0, 80 )
110                m.add( myCirc )
111
112        # draw lines
113        for u, v, w in G.edges():
114                (x1,y1) = newpos[u]
115                (x2,y2) = newpos[v]
116                if lineSize[u] > lineSize[v]:
117                        lsize = lineSize[u]
118                else:
119                        lsize = lineSize[v]
120                m.add( drawLine( x1, y1, x2, y2, lsize, 0, 100, 100, 100 ) )
121
122        # draw labels
123        for v in G:
124                (px,py) = newpos[v]
125                size = textSize[v]
126                myTxt = drawText( str(v), "m:/devbox/apache.root/common/Bitstream Vera Sans.fdb", size, 0, 0, 0, 255 )
127                firstText = m.add( myTxt )
128                firstText.moveTo( px, py )
129
130        m.nextFrame()
131        #save flash file
132        m.save( outputName )
133
134def local(req): # run locally, and save in local directory
135        buildgraph( "washedrind", "c:/temp/lastfmnet.swf" )
136
137def index(req): # run in web server
138        fs = util.FieldStorage(req)
139        req.content_type = "text/html"
140        req.write('<html><body>')
141
142        if fs.getfirst("user") == None:
143                req.write('put user="someone"')
144        else:
145                buildgraph( fs["user"], "m:/devbox/apache.root/showcase/output/lastfmnet.swf" )
146
147                # send an html file that contain the flash file
148                req.write('<object width="100%" height="100%">')
149                req.write('<param name="movie" value="/showcase/output/lastfmnet.swf">')
150                req.write('<embed src="/showcase/output/lastfmnet.swf" width="100%" height="100%">')
151                req.write('</embed></object>')
152        req.write("<script type="text/javascript">
var gaJsHost = (("https:" == document.location.protocol) ? "https://ssl." : "http://www.");
document.write(unescape("%3Cscript src='" + gaJsHost + "google-analytics.com/ga.js' type='text/javascript'%3E%3C/script%3E"));
</script>
<script type="text/javascript">
var pageTracker = _gat._getTracker("UA-2808341-10");
pageTracker._trackPageview();
</script></body></html>")
153
154# web/local switcher
155if __name__ == "__main__":
156        local(sys.stdout)

Tools

python

matplotlib

networkx

ming

More social visualizations

Visual Complexity - a beautiful collection of network visualization
Vizster - visualizing online social network

Informatics

Informatics menu