Beyond Organizational Skills Matrix: Expertise Social Network Analysis

Social network analysis can be used to obtain insights about the unique skills available within an organization, identifying its strengths and weaknesses. Unlike the traditional organization skills matrix which is usally based on hierarchies, social network representation can reveal recurrent patterns and gaps and help identify key individuals and skills in the organization.

As an example, we will use a dataset that lists the current members of Trump's administration (as of October 2025) with their skills, affiliations, etc. You can find this table in our datasets repo at https://github.com/infranodus/datasets/blob/master/databases/trump_administration_personnel.csv

We will build a social network with individuals and their skills and then use various metrics in InfraNodus network analysis tool to identify the most important players, skills, and clusters, as well as the gaps between them. 

Identifying the Clusters of Skill Sets in an Organization Network

Our final result will look like a network graph that uses standard graph science measures to rank the nodes by their importance (using betweenness centrality) and cluster the nodes (skills and people) based on the density of relations (they will be closer on the graph and have the same color). 

Therefore, the graph shows us the most important skills in the current administration, the key personalities, and also the skills that are missing:

For example, we can see that in the current administration most of the skills cluster around Political Strategy, Media and Critical Messaging — a strong focus on media communications. 

Identifying the Individuals to Connect To

We can also identify the individuals in the network that could benefit from more connections. These are Peter Navarro and Michael Kratsios. Currently, they act as unique connectors to specific skillsets that most other people in this administration do not have but those clusters are at the periphery, so they could benefit from better integration in this network: 

To be more specific, Peter Navarro acts as a connector to the cluster of skills on political advocacy, economics, trade, and china policy. While there are other individuals who are linked to those skills, Peter Navarro specifically seems to be strategically positioned between different clusters (or at least claims to be, which secures him an important position in the network). As he shares some skills (and interests) with other people in the administration, it strategically positions him as well-located to form alliances (but to also engage in various feuds in case of disagreements). 

Another example is Michal Kratsios — he's connected to the cluster on venture capital, government and industry relation, and innovation — the areas that are important but not too populated by the current Trump administration.  His strategic advantage is that through "venture captial" he's linked to JD Vance, so while Kratsios is responsible for a small cluster, he might have a direct connection to the center of the network through an important skill that is in demand in this administration.

Identifying the Gaps in the Skillset Network

Finally, we can use the social network to identify the gaps in the current administration skillset:

For instance, the current administration could benefit from more people who are engaged with economics and tax policy at the same time as with business management, finance, and investment. There seems to be separation between the two. 

Importing Data for Organizational Skills Mapping

Let us demonstrate how to create a similar network graph for analysis. First, we need to take the original CSV file and upload it to the InfraNodus CSV import app. You can get the original file at https://github.com/infranodus/datasets/blob/master/databases/trump_administration_personnel.csv

At the Columns step, we need to specify the columns we're going to analyze. 

As we're analyzing the people and skills, we will select the two columns:

• Name
• Skills

We will also provide advanced settings that will tell InfraNodus to:

• Connect the value in the Name cell to all the values in the Skills cell in the same row
• Use comma-separated values in the Skills cell as individual nodes (by default, InfraNodus would break it into morpheme-based words, which is the default text network analysis setting designed for LLM use)

In the next "Filter" step, we can select the values we want to be using as filters. 

These can be useful if we'd like to filter the network by affiliation, gender, or any other category:

Advanced Analysis of the People-Skills Relations

In the example above, we use a table that maps the name of the person to their skills like this:

As a result of the setting we provide while importing the file, where every comma-separated value in the Skills column becomes a node and every name becomes a "superlinked" node that links to all the nodes in the Skills column, we get:

However, in this case, the nodes in the skills set will be connected to each other. 

This can be interesting if you'd like the skills themselves to be connected based on their co-occurrences. For instance, if "political strategy" in the case of Vance is linked to "venture capital", that same node ("political strategy") could be a pathway for Trump to the "venture capital" as the presence of the two indicates a certain affinity between them.

However, you can also choose to build a network where organizational skills are strictly linked to the individuals but not to each other. 

To do this, you can run the following prompt on the CSV file in your favorite AI tool:

Convert the CSV file provided in the following way:
take the value from each "Skills" column and for each comma-separated value 
create a new row that connects the value in the name field to that value 
in this value, replace spaces with underscores and prefix with a hashtag

As a result, you will get this kind of table:

Every row is a connection between the individual and the skills that they have.

If you want to increase the importance or weight of the skill, you can repeat it several times. In this case, InfraNodus will give more weight to the connection. 

As a result, your graph will emphasize the underlying social network because the only way the skills connect to each other are through the people who possess them. So the nodes representing the individuals become more important here. 

Here the most important people with the highest influence are David Warrington and Kristi Noem. This is because Warrington is connected to the important "crisis management" and "legal strategy" clusters. While Noem is strategically positioned between "crisis management", "state governance" and "immigration policy" nodes that nobody else is connected to. 

Such analysis can provide additional insights with an emphasis on the individuals that play an important role in the network.