Which of those indexes/databases should I be using if I want to query over big data using natural language?
First, NLP based search can be executed on top of any engine (APIs are very handy), relational, kv, graph, filesystem .. so that part is totally irrelevant.
Assuming "big data" in this context is still relational data, then any of those systems would suffice, within their own particular tradeoffs and features.
If you're talking about taking some questions and getting graphs from them, ThoughtSpot does a good job.