Arango RX

Inspiration

I have been dabbling in AI for a few years now, but I had never used NetworkX to actually run natural language queries on a database. This seemed like a great opportunity to hone my AI skills while also learning about ArangoDB and various traversal algorithms.

What it does

It allows the user to use natural language to query the ArangoDB. Depending on the type of query, the program will use either AQL or NetworkX (or both!) in order to fulfill the user's request.

How I built it

I used the Jupyter notebook as a starting point and read up on ArangoDB, AQL and NetworkX to learn more about their capabilities and limitations. It took a few dollars deposited into ChatGPT-4o and a LOT of trial and error, but I finally reached a project that I am extremely pleased with.

Challenges I ran into

Wrapping my head around the NetworkX algorithms was difficult, as was getting the prompts to output useful JSON and Python code.

Accomplishments that I'm proud of

Getting the use_both function to finally work! Getting AQL and NetworkX to play nice together was probably the most satisfying point in the entire process.

What I learned

That NoSQL databases do have their place and the data that can be obtained from them is mind blowing. Also, that the power of AI is amazing and we are just scratching the surface of what it can help us unleash with healthcare data.