I have a table calculation returning the median of a column that surpasses the 5,000 row limit by enough that it may not be as accurate to rest on the idea that the 5,000 rows are a representative sample of the larger subset.
Reason I am doing this is to then hide all other fields and display a single digit visualization.
Is there anything I can do to show a more representative Median/Average value of the whole query; larger than the 5,000 row limit?
Hi @Kamani_Charles,
I believe using a measure with type “median” might give you the result you’re looking for here if you’re looking for a single digit visualisation of the median of a set of data. Happy to explain further if needed!
Thanks, Adam
Hi @adstott90,
Yeah, could you explain further?
I believe what Adam suggests is that you add this kind of calculation to your data modelling layer using:
measure: median_order {
type: median
sql: ${sales_price} ;;
value_format_name: usd
}
Then you don’t have to worry about any limits in the front-end data layer
Hit the nail on the head there @Dawid, thanks!