I found out about Neo4j in Mark Jackson's blog and then found this post by a Neo4j employee explaining how to chart your Facebook network of friends step by step. Not all the steps worked for me, which is probably for the better, since I had to learn cypher (the Neo4j querying language) to complete the chart.
I found a few (semi) interesting facts:
- For every direct connection I have there are 10 additional connections between my friends.
- I share over half of my connections with my wife. Though this is not bi-directional, she has many more connections than me. In the picture below I am the red dot in the middle and my wife is the red dot at the far right. Every dot between the two of us is a shared connection, while dots to my left, above and below are only mine connections.
- Most of my fiends are clustered in neat silos. This picture is a network diagram of about half of my Facebook connections. I am colored red, my wife is colored orange and other colored clusters that are not blue represent unique groups (e.g. friends from school, from work, family)
I am not all too interested in researching Facebook fiends. The interesting part is experiencing network technology that's valuable in recommendation engines, fraud detectors and social applications. Next up, I'll compare my Facebook network graph with LinkedIn.
I found a few (semi) interesting facts:
- For every direct connection I have there are 10 additional connections between my friends.
- I share over half of my connections with my wife. Though this is not bi-directional, she has many more connections than me. In the picture below I am the red dot in the middle and my wife is the red dot at the far right. Every dot between the two of us is a shared connection, while dots to my left, above and below are only mine connections.
- Most of my fiends are clustered in neat silos. This picture is a network diagram of about half of my Facebook connections. I am colored red, my wife is colored orange and other colored clusters that are not blue represent unique groups (e.g. friends from school, from work, family)
I am not all too interested in researching Facebook fiends. The interesting part is experiencing network technology that's valuable in recommendation engines, fraud detectors and social applications. Next up, I'll compare my Facebook network graph with LinkedIn.
Comments
Post a Comment