A graph of the social network of data users at Mercy according to self-identified relationships.

Posted from Instagram

Once a quarter, our data management team here at Mercy sends out a survey to co-workers identified as someone who works with data as a core competency for their job. This could be anyone from a billing representative to a biostatician – anyone who spends a substantial amount of their time in Excel, SQL or Oracle manipulating data.

This go round we included a question asking respondents to identify 5 people within the organization they work with and their primary relationship. Respondents could choose one of the following as a possible connection: collaborator, source, recipient or mentor.

Using the free open-source Excel template NodeXL we1 were able to map the over 800 relationships across the organization. This picture is just a sample now that we have the full data set from the survey.

Some of what we might discover:

  • How does the relationships across teams and departments reflect the traversal of data and other information?
  • How does the relationships between people vary across teams? Are some teams more inclusive of people outside of their group in the work they do?
  • Does the number and type of relationships identified show who our influencers are?
  • Is a smaller group of people with stronger connections better for certain projects or initiatives?
  • So if we have over 800 relationships listed on the graph, how many of those listed were people who were sent the survey?
    By association, how many people were on the graph, but not sent the survey?

The list goes on.

Here’s a photo to show how large the print out is of this early draft.


This is the cool stuff I get to work on. It combines my infatuation with social interactions, data visualization and the concept of data mining. Imagine if we did this for the entire IT department – or all of Mercy!

  1. All credit goes to the super talented Paul Boal. I just turned him on to NodeXL and he’s gone places I never imagined.