One of the reasons BillMonk has been growing so well is that friends add friends who add more friends. You have an incentive to add your friends because it makes your lives easier. Being the geeks we are, we wanted to visualize this.
This is just a small part of the BillMonk universe. Each dot is a person; bigger dots mean more recent user activity (the very first two users, Gaurav and Chuck, are the colored dots). Lines with arrows (which you can’t see very well) represent a new user invitiation, and lines without arrows show friendships.
Here are some observations:
- There are some long chains of person to person to person, which is just neat
- There are lots of clusters, which correspond to real-world groups of friends
- These clusters usually start with one person (the seed), but often other friends pass it on within the same group or even spread it to a new circle of friends (it’s easy to see the people who span multiple groups)
- There tend to be two different group patterns:
- One primary billkeeper is entering bills on behalf of the group (we’d guess these are probably roommate situations)
- Everyone is creating transactions with everyone else
- Lots of people are using BillMonk regularly
- One thing this graph doesn’t show you is that people often check out the site, then come back a week later when they have a need for BillMonk
Whenever we do this kind of analysis, we are always very careful to protect user privacy. The first thing we do is to strip all potentially identifying information. (The exception in this case is for ourselves as users).
February 24, 2006 at 12:11 am |
That is sweet. I love graphs!
February 25, 2006 at 6:52 pm |
this is fascinating to look at; what graphing software did you use? it’d also be interesting to see a distribution of the # of connections per person.
February 25, 2006 at 8:35 pm |
Brian – the application we used to create this graphic is “neato”, and it’s part of the free Linux “GraphViz” package (originally authored by AT&T research, I believe). I agree that a distribution of someone’s connectedness would be quite neat to show; another thing we have a great interest in modeling and visualizing is the time dimension, between how long after you are first “infected” by BillMonk that you start to use it, and the rate at which that happens. As time allows it, these will be done.
One reason we did this anaylsis was to get a sense of to how many unconnected social network graphs there are; how many clumps are there? What’s their size distribution? We actually like clumps a lot because if BillMonk remains clumpy as it grows, that means one way to scale could be to partition our DB by clumps (it’s a bad idea to partition a connected graph, but unconnected subgraphs are ok).
March 5, 2008 at 9:37 pm |
the insurance companies don’t want you to know
Information on the life insurance industry
July 27, 2014 at 8:57 am |
advertising, affilate marketing, affilate program, affiliate, affiliate internet marketing, affiliate marketing, affiliate network, affiliate program, bux, cheap advertising, complete offers, cpa affiliate network, cpc,
direct marketing, free money online, get paid for clicking, get paid for clicks, get paid per click, get paid to, get paid
to click, Get Paid to Complete Tasks, how can i make money, how
to make money, internet marketing, internet online marketing, low cost advertising, make money, make money from
home, make money online, make online money, making money online, marketing, marketing online, money, network marketing, on line
marketing, online advertising, online marketing, paid clicks, paid for clicks, paid per click, Paid
To Click Ads, paid to click advertising, Paid To Click Services, pay per click, product advertising, ptc
advertising, social media marketing, social network marketing
March 5, 2015 at 8:39 pm |
It’s actually a great and useful piece of information. I am satisfied that you shared
this useful information with us. Please stay us informed like
this. Thank you for sharing.
November 22, 2017 at 3:56 pm |
I dugg some of you post as I cerebrated they were handy very helpful