January 20, 2012

InMaps from LinkedInLabs - Data Scientists in action

"It ain't what you know, it's who you know."  Agree?  Disagree?  

However nuanced your response, I believe we all know that relationships matter.  We are creatures that love to communicate.  We need others.  We don't generally achieve greatness in isolation.  

Systems are dynamic, purposeful orchestrations of people, processes, and platforms.  Systems are usually  multi-dimensional beasties that don't like having their picture taken.  We illustrate systems using "views" that emphasize its different aspects.  Like the blind men and the elephant, we can tend to characterize the system according to the view that we most directly use.  A set of views enable us to build up a wholistic understanding of the system by looking at it from different perspectives.  

The organizational chart is a hugely important view of your system.  Your organizational chart focuses on the People view of the system and highlights authority and influence structures.  As a time-tested and well-loved view, the organizational chart works very well in an established hierarchical context.  

But, what if your relational network doesn't fit nicely into the abstract boxes and lines of an organizational chart?  What if you have relationships that cross professional, corporate, and physical boundaries?  How do you find a view of your system of relationships?  How do you find people with influence?  How do you keep important relationships from dropping off the map?

My answer (right now) is InMaps.

InMaps is provided by the LinkedIn Analytics team. You can check out the official video, brought to you by DJ Patil, LinkedIn's chief scientist (one of the creators of the data scientist role) and published on the LinkedIn Analytics blog.   InMaps provides a powerful illustration of the People view.  

InMaps visualizes all of the people that you know, connecting them to all of the people that you mutually know.  This creates your relational map.  As LinkedIn calls it, this is your professional universe.  

I use this view of my relationships to:

1. Monitor change in my professional network
2. Understand the "people" environment
3. Focus networking and team-building
4. Identify and navigate opportunities
5. Keep in touch with old friends

Here is a recent screenshot of my map:

I can click on anyone in this map, and view the network of people connections that we share.  In this example, I have clicked on my brother, David.  

David is a hugely talented graphic artist and web designer.  Check out his work and get to know him at SMRVL.  

As you can see, David and I have a lot of overlap within the blue network.  We share a few family members in my purple network.  Other than that, David doesn't know too many folks in my other networks.  

David was kind enough to share his map with me.  (One major -- though necessary -- limitation is that you can only see your own map.)  Here is his full map.  


Here's what David says about his map.  "I like it because it's a brilliant at-a-glance snapshot of my web of relationships. I can, at a glance, see where I have relational density ... the two big masses on my map are both social networks."

This looks great.  But in order to get anywhere, we need to know how to read this map.  Here are some questions that I ask my map in order to find meaningful stories.

1. How many "lobes" do I have?  InMaps will automatically color-code clusters that it thinks are distinct networks.  The more good-sized lobes that I have on my map, the more valuable I think my network is.  For me, this highlights a relational victometric -- Number of Significant Bridged Networks.  I want to bridge 3 or more significant professional networks in order to minimize risk as an independent contractor.  Connections to multiple networks represent opportunities to make things happen.

2. Who are the "foci"?  Who are the people who are well-connected within communities?  They are the people who know people.  When I'm seeking to get better connected with a group, I'll try to start with these gregarious types.  

3. Who are the "edges"?  Who are the people who seem to be minimally connected?  Here, I might read more closely.  If the person is someone whom I know to be outgoing, they may connect me with new networks to explore.  Many people simply don't use LinkedIn that much.  The loosely connected person may be someone who would be open to learning more about what their tools can do for them.

How do you get your own map?  Building your map is easy.  If you have a LinkedIn account, go to http://inmaps.linkedinlabs.com/ and they'll walk you through the steps from there.  You'll need to decide how you want to share access to your data.  I chose to share for a month, and have just continued to allow access.



A couple of extra factoids:
  • Originally, InMaps needed you to have at least 50 connections in order to build your map.  I'm not sure if that's still true.  If it is and you have less than 50 people in your relational network, get busy!
  • InMaps are updated every Sunday.  If you add to your network, don't expect them to show up until sometime on Monday.  I usually capture a screenshot every week.  I dump these into PowerPoint.  As I play through the slides chronologically, I can more easily highlight changes and realignments in my network.
  • On the splash screen, hit "Next Map" in the lower-right corner to see different shapes of networks.
  • InMaps uses Hadoop/Pig , Ruby , VoldemortJava and Processing for the heavy backend maps calculation and image processing.
  • For more, check out the very informative InMaps FAQ.