By now everyone is aware of the massive boom in social-networking (Twitter, Facebook, LinkedIn) and obviously a big part of its business model revolves around being able to mine this data to create information that can be used to make money for someone.

Gartner has identified 'Social Analytics' as one of the top 10 strategic technologies for 2011.

  • Has anyone looked at their existing data structures to determine if they could extract a social graph and then perform further data mining against this?
  • How does it fit in with your other strategic development strategies?
  • What information are you trying to extract from the data?

Take for example, a bank. They could conceivably determine a social graph through account relationships and transactions. Obviously there would be open edges on the graph where funds enter/leave the institute, but that shouldn't detract from the usefulness of the data.

I'm looking for actual examples with the answers, as well as why/how they did it. References to other sites will be greatly appreciated.

Note: I'm not at all referring to mining data out of actual social networks.

  • I am not sure that it would even be legal for a bank to do something like you suggested.
    – Pemdas
    Jan 9, 2011 at 4:22
  • @Pemdas That is just an example; but why would generating reports for themselves, from their own data be illegal? I'm sure if they ever did anything like that, it would have to be cleared by their in-house legal departments anyway... Jan 9, 2011 at 4:25
  • I don't know, I would be interested in understanding the privacy issues around something like this. I really don't know the answer, but facebook clearly has a wealth of data that is public.
    – Pemdas
    Jan 9, 2011 at 4:34
  • Agreed. Not that I really know them, but I have the feeling current laws are far behind the new capabilities we have with analysing data. Jan 9, 2011 at 4:36
  • From the lack of answers, it would appear nobody does. Jan 10, 2011 at 11:01

1 Answer 1


I worked on a set a internal business applications that had a smaller set of users (about 100). The applications were flexible tools that might have different uses depending on the users objectives. We tracked many user actions and stored them in a database. Basically when a user committed an action that required retrieving or update the database we added a log for that activity.

The analysis on the data has mostly been grouping people together based on how they used the tool and looking to see if they were in the same department and or role. This allowed us to target new features to what we saw people were trying to do and helped us understand areas that departments might have overlap.

  • Not what I had in mind, but only because I hadn't considered that line of thought during the question. That sounded quite interesting and I'm sure we have some internal data we could extrapolate this sort of data out of. +1. Jan 9, 2011 at 5:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.