Make average work as in pandas

Author: jecisc

src/DataFrame-Math/DataPearsonCorrelationMethod.class.st

@@ -11,8 +11,8 @@ DataPearsonCorrelationMethod class >> between: x and: y [
 	| xDeviation yDeviation |
 	x size = y size ifFalse: [ SizeMismatch signal: 'Correlation can not be calculated for two series of different size' ].
 
-	xDeviation := (x values replaceAll: nil with: 0) - x average.
-	yDeviation := (y values replaceAll: nil with: 0) - y average.
+	xDeviation := (x values replaceAll: nil with: 0) - (x sum / x size).
+	yDeviation := (y values replaceAll: nil with: 0) - (y sum / y size).
 
 	^ (xDeviation * yDeviation) sum / ((xDeviation ** 2) sum * (yDeviation ** 2) sum) sqrt
 ]

src/DataFrame-Tests/DataFrameStatsTest.class.st

@@ -32,6 +32,23 @@ DataFrameStatsTest >> testAverage [
 	self assert: actual closeTo: expected
 ]
 
+{ #category : #tests }
+DataFrameStatsTest >> testAverageWithNils [
+
+	| expected actual |
+	df := DataFrame withRows: #( #( 1 1 ) #( 2 nil ) #( nil 3 ) #( 4 4 ) ).
+	df columnNames: #( dogs cats ).
+
+	expected := {
+		            (7 / 3).
+		            (8 / 3) } asDataSeries.
+	expected name: #average.
+	expected keys: df columnNames.
+
+	actual := df average.
+	self assert: actual closeTo: expected
+]
+
 { #category : #tests }
 DataFrameStatsTest >> testCorrelationMatrix [
 

src/DataFrame-Tests/DataSeriesTest.class.st

@@ -531,6 +531,18 @@ DataSeriesTest >> testAtTransformIfAbsent [
 	self assert: exceptionBlockEvaluated
 ]
 
+{ #category : #'tests - arithmetic' }
+DataSeriesTest >> testAverage [
+
+	self assert: #( 1 2 3 4 ) asDataSeries average equals: 5 / 2
+]
+
+{ #category : #'tests - arithmetic' }
+DataSeriesTest >> testAverageWithNils [
+
+	self assert: #( 2 nil 4 ) asDataSeries average equals: 3
+]
+
 { #category : #'tests - comparing' }
 DataSeriesTest >> testBooleanGreaterThanEqualFromScalar [
 

src/DataFrame/DataSeries.class.st

@@ -160,6 +160,13 @@ DataSeries >> atIndex: aNumber transform: aBlock [
 	self at: key transform: aBlock
 ]
 
+{ #category : #information }
+DataSeries >> average [
+	"We do not count the nils"
+
+	^ (self values reject: #isNil) average
+]
+
 { #category : #'data-types' }
 DataSeries >> calculateDataType [