Mapping Strategies For Survey Data

Elastic Stack Logstash
1

I am ingesting data and looking for opinions on better ways of indexing the data for later aggregation analytics. The data is currently pulled and stored in Elasticsearch as numbered q&a fields.

{
  "question1": {
    "question": "How are you?",
    "answer": "Pumpkin Pie"
  }
},
{
  "question2": {
    "question": "What's your favorite color?",
    "answer": "Bowling ball"
  }
}

The problems I am seeing with this strategy is, there's no easy way to create a Kibana visualization, that I know of, that aggregates all the questions and/or answers to answer things like "How many times is the question, 'How are you' asked?" Or a way to filter on a specific question without having to know it's 'question246'. I thought maybe I could join all the questions/answers asked into a single field of arrays, but I'm not sure how I'd go about joining a non-contiguous N number of fields like that together via Logstash.

Any thoughts on a better way to organize the data?

2

How about making each question its own event with [question] and [answer] fields?

3

heh, this is actually related to another post I recently opened that you helped me with. The questions asked are actually part of an event and need to stay with that event.