Recommendation engine based on Employee Data

Node Link


Has anyone tried to build a recommendation engine based on employee info?

Employee info may consist of Name, Email ID, Manager, Location, Skills, Interest, Emailed to recipients, emails from recipients, Events attended etc.
Basically, we are trying to mine employee data and all the activities and create a recommendation engine to suggest things like "Employees you may know" this will be based on similar skills, interest, events participated, may be location etc. or "Training/course you may be interested" based on skills data or "Event you may be interested" based on employees Interest data. The scope is endless.

I recently took some neo4j training, so just have some basic knowledge of loading data, writing cypher queries to create nodes, labels, relationships, properties etc.
I am aware, this use case requires advance knowledge on programming skills and graph algorithms, It might also require some ML skills, NLP knowledge etc. I do have basic understanding of those skills.

It would be great to know if someone has done something similar and if they can guide me in the right direction. Or have an idea as to what steps are required to make this possible. Eg.
What all algorithms can be used (Similarity, Communities etc.)? What additional plugins/applications required such as GraphAware, Graph Analytics etc.


  • I am using Neo4j 3.5.14 version along with Neo4j Bloom 1.2.0
  • Applications installed: Graph Apps Gallery, Neo4j etl tool, Graph Gallery, graphxr, Graphlytic desktop
  • Plugins: APOC and Graph Algorithms

Excited to hear your ideas! Thank you!!!


Graph Voyager

I would start with the graph algorithms plugin like you've mentioned, there are both community detection and similarity algorithms that could help.

Check out the free graph algorithms ebook for some examples and walkthroughs, your use case is similar enough to those to get you going (or lead you to more specific questions). Good luck!

Node Link

Hi Pooja

You may find this blog post interesting:

At the end of the post, there is a also link to a GraphGist.