Stacked bar chart

The stacked bar chart widget breaks each bar into segments, so you can see both the total and how it splits across several series. It is ideal for composition — how the parts of a whole change across a category or over time.

Use it for:

• Platform share per day (Steam vs iOS vs Android stacked)
• Event volume split by app version over time
• Revenue composition by product line per month

Note

Use a stacked bar chart when the segments add up to a meaningful total. If you only care about comparing categories independently, use a bar chart; for trends without stacking, use a line chart.

---

Query requirements

Return one label column (the X axis — a category or time bucket) followed by one numeric column per stacked series. Each numeric column becomes a stacked segment, and its column name appears in the legend.

The series must be known in advance, so use conditional aggregation (countIf, sumIf, uniqExactIf) — one expression per series.

Simple mode (recommended)

Select

toDate(timestamp)                       AS date,
countIf(platform = 'steam')             AS "Steam",
countIf(platform = 'ios')               AS "iOS",
countIf(platform = 'android')           AS "Android"

Filters

WHERE name = 'app_started'
GROUP BY date
ORDER BY date

Result shape:

date	Steam	iOS	Android
2026-04-01	120	85	43
2026-04-02	134	91	38

Each row stacks its Steam, iOS, and Android values into a single bar.

---

Display format

Set Display format to percent to turn each bar into a 100% stacked bar — every segment is shown as its share of that bar’s total. Return raw counts; the normalization is handled for you.

---

Common examples

Platform composition per day

Select:

toDate(timestamp)               AS date,
countIf(platform = 'steam')     AS "Steam",
countIf(platform = 'ios')       AS "iOS",
countIf(platform = 'android')   AS "Android"

Filters: WHERE name = 'app_started' GROUP BY date ORDER BY date

---

Sessions by version per week

Select:

toStartOfWeek(timestamp)                        AS week,
uniqExactIf(session_id, app_version = '1.7.9')  AS "v1.7.9",
uniqExactIf(session_id, app_version = '1.7.8')  AS "v1.7.8"

Filters: WHERE name = 'app_started' GROUP BY week ORDER BY week