Scatter plot

The scatter plot widget plots each row as a point positioned by two numeric measures — one on the X axis, one on the Y axis. It is the tool for spotting correlation, clusters, and outliers across your users, sessions, or events.

Use it for:

• Sessions per user vs total playtime
• Level reached vs money spent
• Any two per-entity metrics you suspect are related

---

Query requirements

Return rows with at least two numeric columns:

• The first numeric column is the X axis.
• The second numeric column is the Y axis.
• An optional non-numeric column is used as the point label (shown in the tooltip).
• An optional third numeric column drives the bubble size.

Each row becomes one plotted point, so aggregate per entity (e.g. GROUP BY identity).

Sessions vs playtime per user (full SQL)

SELECT
  identity,
  uniqExact(session_id)                                       AS sessions,
  dateDiff('minute', min(timestamp), max(timestamp))          AS playtime
FROM {table}
WHERE name = 'app_started'
GROUP BY identity
LIMIT 1000

Result shape:

identity	sessions	playtime
a1b2…	12	340
c3d4…	3	45

Here sessions is the X axis, playtime is the Y axis, and identity labels each point.

Note

Scatter plots can render many points. Add a LIMIT to keep the chart responsive, or pre-aggregate to one point per entity.

---

Bubble size

Add a third numeric column to scale each point by a third dimension:

Full SQL:

SELECT
  identity,
  uniqExact(session_id)              AS sessions,
  count()                            AS events,
  countIf(name = 'purchase')         AS purchases
FROM {table}
GROUP BY identity
LIMIT 1000

sessions → X, events → Y, purchases → bubble size.

---

Common examples

Engagement vs spend per user

Full SQL:

SELECT
  identity,
  count()                                          AS events,
  sumIf(toFloat64(value.amount), name = 'purchase') AS spend
FROM {table}
GROUP BY identity
LIMIT 1000