Making the move to graph from SQL - Looking for patterns in repairs

[steve_webb]

Hi, I am based in the UK and only just started the journey into Graph DB's
My current role is technical support manager and I hope to use graphs to look for patterns in support cases raised and also for repairs, common component failures.
At this stage, I am not even sure this is a good fit.
I have produced a few MS Power BI dashboards for people to use but need to get smarter.
Ultimately using Machine Learning to allow the faster resolution to common problems submitted to the support desk.
Any tips would be gratefully appreciated.

[clem]

The huge value in Graph DB's is that they can model many-to-many relationships, unlike traditional "Relational" Databases.

I think your use case does make sense because it's likely your data has a lot of these many-to-many relationships.

I imagine you might have a case where a component is made up of several subcomponents that are shared among other components. And the components can go into multiple products.

Subcomponents could in turn come from multiple suppliers, some of whom are less reliable than others.

And of course, your customers may have multiple different products of yours at multiple locations.

Finally, you're getting started. You may discover that you need to take into account of data that you hadn't thought of or wasn't previously collected. Neo4J is flexible, so adding new stuff to your DB is painless.

The various APOC functions can help you modify your DB in case you need to rename Labels, etc.

I also find the tutorial and documentation for Neo4J to be above average, so I think you'll find it easier to get something going than with other DB's.

Good luck!