Is there a was to use graduated symbology and have it calculate and classify the data by percentile?
I'd like to group my data in the following percentiles:
0-25 %
25-50 %
50-85 %
85-100 %
-
See also here: gis.stackexchange.com/q/470726/88814– BabelCommented Nov 22, 2023 at 21:29
Add a comment
|
2 Answers
You can use Categorized styling with an expression.
My field with values is named field1
CASE
WHEN "field1"< (25*( count( "field1")+1))/100 THEN '0-25 %'
WHEN "field1" BETWEEN (25*( count( "field1")+1))/100 AND (50*( count( "field1")+1))/100 THEN '25-50 %'
WHEN "field1" BETWEEN (50*( count( "field1")+1))/100 AND (85*( count( "field1")+1))/100 THEN '50-85 %'
WHEN "field1" BETWEEN (85*( count( "field1")+1))/100 AND (100*( count( "field1")+1))/100 THEN '85-100 %'
ELSE 'ERROR'
END
-
This is an elegant solution but I think it applies only to one special case: when "Field1" contains a set of inverse-ranked values from 1 to n (n being the number of records, n being largest rank and 1 being smallest). In any other case, one can imagine situations where all values fell into the "0-25%" category or that some or all fall into the "ERROR" category (e.g assuming 100 records, any value above 101 would be so). It would be awesome if this could be generalized to apply to any set of numbers by somehow doing the ranking first or using percentiles as the case limits. Commented Aug 11, 2023 at 22:10
Here is a generalized solution inspired by (aka largely stolen from) this answer by @MrXSquared here: Make "quantile" classification with an expression
with_variable(
'percentile',
array_find(array_agg("field1",order_by:="field1"),"field1") / array_length(array_agg("field1")),
CASE
WHEN @percentile <= 0.25 THEN '0-25%'
WHEN @percentile <= 0.5 THEN '25-50%'
WHEN @percentile <= 0.75 THEN '50-75%'
WHEN @percentile <= 1 THEN '75-100%'
ELSE 'Undefined'
END
)
