Luma and my university wont give me the data for me to run a community aggregate plugin, so I recreated it to enable my plugin idea

Its time our event hosting platforms let us perform analytics across events to grow our communities.

I've sunset the domain and services to save on cost while I build an MCP related project. Feel free to use this codebase.

Frontend: NextJS TailwindCSS

BE: ExpressJS fastAPI microService OpenAI LLM Calls Supabase

NextJS : I'm comfortable with this so its more from a comfort perspective alongside not having to worry about dynamic routing

TailwindCSS : Used for Shadcn components and love the way I can write inline CSS

ExpressJS : Mainly for speed in doing concurrent events (most of the time people are creating events or joining them, the event analytics is for the hosters). When building past MVP will allow me to handle blasting out emails and or other concurrent applications with ease.

fastAPI : run my ML computations in a python microservice since python has more developed access to models such as XGBoosting for tabular data and I'll leverage it to do my LLM calls.

This current quick mockup does NOT involve the use of customizable event questions. Main reasoning is for any decent XGBoosting Model, I require substaintial data that reflects the events needed to collate such probabilities. While I could do custom trained for each and build alongside, the variance for the first few events (which in this case time period would be > a year) is too high to consider alternatives without heavy theoretical ML knowledge. Since I lack that as of writing this documentation I have taken the step instead to be be a focus on imagining what it could be until sufficient dataset can be used to properly train a model.

• Make profile pictures fully customizable
• Customizable calendar views + more precise XGBoosting models (requires data)
• Allowance for custom event logic and tabling (This implies a more robust way of knowing and developing trad models, LLMs aren't harmed)