Project — Analytics Dashboard

Build the mental model

Lesson 11 introduced aggregation basics (metric and bucket), and this project combines several aggregations, nested together, into a single request to build a real analytics dashboard — nesting a `date_histogram` bucket aggregation (time-based buckets, say "how many events occurred per day") inside a `terms` aggregation gets you two-dimensional analytics — "view count per day, per topic" — from a single query, serving the same purpose as SQL's `GROUP BY topic, date_trunc('day', occurred_at)` with two levels of grouping. Kibana was introduced in Lesson 7 as the Dev Tools console, but its core purpose is visualization — instead of manually parsing an Elasticsearch aggregation query's JSON response and drawing a chart yourself, Kibana's "Lens"/"Visualize" UI lets you point at an index pattern and interactively build bar charts, line charts, and data tables via drag-and-drop, with Kibana itself auto-generating the underlying aggregation query. To build a production-grade analytics dashboard, store raw event data (like each user's lesson-view event) in a separate time-series index (applying Lesson 16's ILM concept here) rather than mixing it with the `tutorials` catalog index — analytics data's growth pattern (daily events) and catalog data's growth pattern (occasional edits) are so different that they need separate index and lifecycle strategies.

Connect it to a real scenario

From the Tutorial Platform's `lesson-views` event index (documents shaped `{ tutorialId, tags, occurredAt }`, with a Lesson 16-style ILM policy), write a nested aggregation query for "view count per topic (tags), per week" — nesting a `date_histogram` (weekly interval) inside a `terms` aggregation (on `tags`). Prototype this query in Kibana's Dev Tools first, then switch to Kibana Lens and build a "most-viewed topics this month" bar chart and a "weekly completion trend" line chart with drag-and-drop, pinning them into a dashboard to share with the admin team.

Nested aggregations (terms + date_histogram) serve the same purpose as a SQL multi-level GROUP BY
Kibana Lens/Visualize auto-generates aggregation queries and builds interactive charts
Keep analytics event data on a separate index and lifecycle strategy from catalog data

Try the working example

http

GET /lesson-views/_search
{
  "size": 0,
  "aggs": {
    "by_topic": {
      "terms": { "field": "tags", "size": 10 },
      "aggs": {
        "by_week": {
          "date_histogram": { "field": "occurredAt", "calendar_interval": "week" }
        }
      }
    }
  }
}

You should see

You get nested buckets of weekly view counts under each tag — ready to chart in Kibana Lens.

5-minute try-it

Write a query for "average views per difficulty level" by nesting an `avg` metric aggregation inside a `terms` (difficulty) aggregation — describe how you would configure the result as a Kibana bar chart.

One important caution

Storing high-volume analytics event data (daily user activity logs) inside the same `tutorials` catalog index — analytics data disrupts the catalog index's size and growth pattern and can hurt search performance too.

Leaving the outer `terms` aggregation's `size` parameter at its default (10) inside a nested aggregation — if topic count exceeds 10, some buckets are silently dropped and the chart's data ends up incomplete.

Kibana Guide — Create a Visualization — Elastic