What is the fundamental difference between analytics in Kibana and OLAP cubes?

Question

OLAP

Greetings.
I'm rummaging through information on business intelligence, trying to understand the subject, and is it possible to use elasticsearch + kibana as an olap cube engine.

Answer 1

I did not work with OLAP cubes and specific software, but I got to know it superficially. It looks very confusing and enterprise, I'll try to stay away from them :)
Regarding analytics using ES - it's possible, I did it - the impressions are positive. To do this, it is enough to determine what type of documents in your database is considered central, after which, when adding, perform recursive denormalization of data - nested documents will help with this. You should get something along the lines of:

"offer": { // данные из таблицы товаров
  "price": {
    "value": 1500000,
    "currency": "RUB"
  },
  "category": { // данные из таблицы категорий
    "name": "Иномарки",
    "group": { // еще одни данные из таблицы категорий
      "name": "Автомобили"
    }
  },
  "owner": { // данные из таблицы пользователей
    "first_name": "Антон",
    "last_name": "Антонов",
    "locality": { // данные из таблицы городов
      "name": "Москва",
      "country": { // данные из таблицы стран
        "name": "Россия"
      }
    },
    "organization": { // данные из таблицы организаций
      "name": "FreeAuto",
      "phone": "[phone]"
    }
  }
}

Further selection from the resulting index is made by aggregations, which can be multiple, nested, filtered, by a specific field, by a specific field in a nested document, etc. Here it depends on your tasks what and how you will aggregate.
And this, Kibana, is nothing more than a beautiful interface to the results of aggregations; nothing at all depends on it.

Answer 2

It was the same thought. Only I have the opposite: I worked quite tightly with OLAP. but with Elasticsearch + Kibana only as a user. Didn't notice any particular differences. In terms of the final result - BI.
The differences may be in the understanding of OLAP itself. When it looks not like a visual graph, but like a table with the ability to cut and regroup data in real time. In the micro version, you can play with this in Excel or Google Sheets. There it is called Pivot Tables. In OLAP, in terms of UI, you can only process billions of rows and store terabytes of data.