Hello Baserow team,
Thanks for the existing formula aggregations (min, max, avg). One addition would unlock more robust analytics: median() (and, if possible, percentile(array, p)).
Why median?
Median is resistant to outliers and better reflects “typical” values in skewed data (e.g., response times, resolution times, order values).
Example (response time, minutes):
-
10× 5 min → 50
-
12× 7 min → 84
-
1× 1000 min → 1000
Total: 1,134 minutes over 23 responses → avg = 49 min
Median of the sorted list is 7 min, which better matches user experience.
Where it would help:
-
SLAs/Support: response or resolution time medians (YTD, by team, by channel)
-
Product Ops: median cycle time per feature
-
Finance: median order value when the distribution is skewed
-
HR: median time-to-hire, etc.
Suggested scope:
-
Formula function:
median(array<number>) -
Rollups/group footers: show Median alongside Min/Max/Avg
-
(Optional)
percentile(array<number>, 0.9)for p90/p95
Definition notes:
-
For even counts, return the average of the two middle values.
-
Ignore blanks/nulls by default (consistent with other aggregations).
Is this on the roadmap? I’d be happy to share sample datasets/KPIs if helpful.
Best,
Felix
