Formula for detecting local (contextual) outliers in DAX
I have created a calculated column that looks for global outliers but have found that the be insufficient as I have some data in there that, over the entire dataset, is ok but it is obviously incorrect if you look at the values around it. I can't find a simple formula for this that I can use in DAX. Does anyone know how I could modify the DAX below so it looked at the surrounding values (excluding values that are already marked as a global outlier) and marked any anomilies? = var Q1 = CALCULATE ( PERCENTILE.INC ( 'sampleDataForCalcs'[Machine Hours - Raw],0.25 ), FILTER ( ALLEXCEPT ( 'sampleDataForCalcs', 'sampleDataForCalcs'[Unit Number] ), 'sampleDataForCalcs'[Machine Hours - Raw] > 0 ) ) var Q3 = CALCULATE ( PERCENTILE.INC ( 'sampleDataForCalcs'[Machine Hours - Raw],0.75 ), FILTER ( ALLEXCEPT ( 'sampleDataForCalcs', 'sampleDataForCalcs'[Unit Number] ), 'sampleDataForCalcs'[Machine Hours - Raw] > 0 ) ) var IQR = Q3 - Q1 var low = Q1 - 1.5*IQR var high = Q3 + 1.5*IQR var outlier = IF ( 'sampleDataForCalcs'[Machine Hours - Raw] < low || 'sampleDataForCalcs'[Machine Hours - Raw] > high, "TRUE", BLANK() ) return outlier
