I am trying to make an equal interval legend using a rate of burglaries per 1000 habitants using a combination of a column with the counts of burglaries and a column of total population using census districts. Due to the fact that there are outliers in the data (extremely high values) I would like to exclude them from the process of defining classes in the legend based on a threshold of the total population (less than 700). I already build a expression in the expression function (which exclude values <700) window like this:

CASE WHEN "burglaries_CS_POPN_2000|2000" >= 700 THEN ( "burglaries_P500_ALL|2009" / "burglaries_CS_POPN_2000|2000" ) END

It works fine, but the polygons excluded are not showed anymore. In ArcGis there is an option for "show symbol for excluded data" in the symbology-classification window. Is there an option in QGIS that could perform the same task without adding a new column in the attribute table? Could it be done using an expression like I show above?

1 Answer 1


I actually will remove values above a certain threshold (1.5 Standard deviations above and below the average and rerun interpolation sometimes to show the more subtle differences in the data. I will make one layer that has the outliers still in it with a color scheme like shades of red. Then I will place a transparent outliers-removed version over it so that the super high readings show up below while still displaying the more subtle (though sometimes more important) clouds or clusters.

I would also like to see some better option!

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.