IKANGAI Solutions. e.u.

After the initial analysis that included the basic network stats like centrality indices or network density, we took a more detailed look at the social network data. Our first interest was how often our advertisement Tweet was retweeted. We used the Twitter API to get a list of Twitter users that retweeted:

http://api.twitter.com/1/statuses/181034348781383683/retweeted_by.xml?count=100

In this case, the Tweet was retweeted at least 100 times over (Twitter does not provide detailed information on the exact number), which we believe is quite a success for us . We took a first sample of the Twitter users that retweeted the Tweet from meghannz96. Surprisingly, none of the sampled Twitter users could be found in the follower network of meghannz96. We will elaborate on this and check, if we can find a overlay. We also looked for communities and we found six different communities (denoted by different colors) as shown below.

community structure of meghannz96 Twitter ego network

Finally, we checked the impact of the Tweets on our downloads: after the Tweets were published, we noticed an increase of downloads to 35 downloads per day. Previously, we had between 15 and 20 downloads per day. The traffic on our Web page also increased and increased to 903 hits for the product Web page http://www.ikangai.com//apps/qlauncher/ that was included in the advertisement Tweets. This is very interesting, since we had virtually no hits for this particular Web page in the months before and the advertisement Tweets have apparently a considerable impact.

your ikangai science team

In an previous experiment, we bought 1000 Twitter followers and analyzed the resulting network. This time, we experimented with marketing on Twitter. During the last few days, we used the services of Twitter users that offer to buy Tweets for advertisement purposes.
After our advertisement Tweets were posted, we decided to take a look at the ego networks of one of our Twitter users that was tweeting for us.
We collected the network data of a Twitter ego network with Node XL. After quickly looking at the data (see below), we can say that the ego network is (not surprisingly) centered around a single ego (meghannz96). Paths in the network are short: basically, we can reach every node in the network with two hops (the average geodesic distance is almost two). The density of the network is rather low with the network having about 1% of all possible links. The distribution of the in and out degree appears to follow a power law distribution which would make the network scale free.

Connected Components: 1
Single-Vertex Connected Components: 0
Maximum Vertices in a Connected Component: 1654
Maximum Edges in a Connected Component: 26307
Maximum Geodesic Distance: 2
Average Geodesic Distance: 1.988184
Graph Density: 0.009621947
Minimum Out-Degree: 0
Maximum Out-Degree: 1483
Average Out-Degree: 15.905
Median Out-Degree: 6.000
Minimum Betweenness Centrality: 0.000
Maximum Betweenness Centrality: 2307580.543
Average Betweenness Centrality: 1635.456
Median Betweenness Centrality: 4.667
Minimum Closeness Centrality: 0.000
Maximum Closeness Centrality: 0.001
Average Closeness Centrality: 0.000
Median Closeness Centrality: 0.000
Minimum Eigenvector Centrality: 0.000
Maximum Eigenvector Centrality: 0.015
Average Eigenvector Centrality: 0.001
Median Eigenvector Centrality: 0.000
Minimum Clustering Coefficient: 0.000
Maximum Clustering Coefficient: 1.000
Average Clustering Coefficient: 0.333
Median Clustering Coefficient: 0.264

The ego network of the twitter user meghannz96 is plotted below.

meghannz96 ego network

For comparison, we selected a random node and plotted the resulting sub-network:

sboysuccess sub network

There is a lot of additional information in our dataset (e.g. number of tweets or the timezone of the twitter account) which we will analyze and post in the next few days.

your ikangai analysis team