Suppose you want to build a graph of your business network where you link people and companies they work with, as well as the professional relations between the people, their skills, areas of expertise, and the problems they are working on.
You can use InfraNodus to create this map and the resulting graph can be used to both have an overview of your organizational structure as well as to find the structural gaps in it in order to optimize the processes within.
Additionally you can then plug this graph into your favorite LLM like Claude using the InfraNodus MCP server in order to have this knowledge available to your AI agents and in your AI workflow.
Here is an example of how you build such graph step by step. There are generally two ways to do that:
Using a spreadsheet, which you then import as a CSV table
Manually using plain text input
In both cases, you should be using [[wikilinks]] syntax to designate nodes to avoid automatic lemmatization in InfraNodus. When you import a CSV file you can add [[wikilinks]] automatically to the columns you're adding, but if you're creating a spreadsheet from scratch, it's better to use them from the beginning.
We provide the two options below, feel free to copy and paste the instructions to your favorite AI to use as a prompt to automatically create the CSV file based on your own data.
Option 1: Building a Network from a Spreadsheet
One way is to just create a spreadsheet for each type of graph: social relations, skills, problems people are working on, companies. Then you import each CSV into a separate graph and you can query them separately or merge them together.
Step 1: Creating a Spreadsheet CSV File
< syntax instructions (to use as AI prompt also) >
For instance, for the graph of the social relations, you can make a spreadsheet that looks like this (CSV format):
[[Bogdan]],[[Dmitry]]
[[Bogdan]],[[Emma]]
[[Dmitry]],[[Anatoly]]
Another graph could be for what they’re working on. If one person has several skills — you should add each skill as a new row:
[[Bogdan]],[[Organization Mapping]]
[[Bogdan]],[[Network Science]]
[[Dmitry]],[[Software Development]]
[[Emma]],[[Law]]
Etc
You can add an extra column for the type of relationship (e.g. “knows” or “is specialist in”).
< end of syntax instructions >
Step 2: Importing the Data
When you're importing this data, you'll choose these columns as the columns to import:
If you have the type of relation column, add this column as a filter column on CSV import:
Then you import the files into two separate graphs, so you have a graph of the social network and another graph of the skills the people in the organisation have.
Then you can use the “Compare Graphs” feature to merge them together
Option 2: Building a Social Network Manually via Text Graph Input
Here you simply create a new graph and then start adding nodes and edges via the Content Editor + Add Content function.
If you're using the InfraNodus MCP server plugged into your Claude or Codex, you can give it the instruction below to create a graph from your raw data:
< syntax instructions >
Use the [[wikilinks]] syntax to designate entities and a statement to describe the relations between them.
For example, for a social network graph:
[[Bogdan]] knows [[Dmitry]]
[[Bogdan]] works with [[Emma]]
For a graph of what people are working on:
[[Bogdan]] does [[Organization Mapping]]
[[Bogdan]] is a specialist in [[Network Science]]
Optionally, if you'd like to add a type of relation, you can do it via tags in the text input (or adding tags to each statement):
That's it, this is how you build the graph and then you can use advanced network science metrics to identify the key nodes, clusters, and gaps between them.
Comments
0 comments
Please sign in to leave a comment.