SportsBot Training Data and Modal

Building out SportsBot smart entity and looking for any good public or private sports data sources to train our sports focused chat agent with a multimodal LLM. Also looking for any input on which llm to use (imagebind,etc) and why… We hope to expand sports knowledge and education with a give back mission using Sportsbot.com in a DAO structure to execute the conversation and entity. Been researching for weeks and hoping to get any input or help… Much appreciated.

I am fairly to the LLM world, so I apologize if this is a beginner question and I am in the wrong location…
Respectfully, edad

We have some high school and d3 sports data available.