A Simple Linear Regression function for M / Power Query

fnSimpleLinearRegression.M

// Found at https://www.bookkempt.com/2017/10/simple-linear-regression-in-power-query.html
// Expects a table with numeric columns x and y.
// To run over an entire table, first Group By nothing and Aggreagate All Rows. 
// To run over subsets of a table, Group By columns defining your groups and Aggreagate All Rows.
// Then use Add Column / Invoke Custom Function and pass the All Rows column as the myTable parameter.
// Expand the resulting new column to get Slope, Intercept and RSquared

(myTable)=>
let
    AddedXY = Table.AddColumn(myTable, "xy", each [x]*[y]),
    AddedXX = Table.AddColumn(AddedXY, "xx", each [x]*[x]),
    CountX = List.Count(AddedXX[x]),
    SumX = List.Sum(AddedXX[x]),
    SumY = List.Sum(AddedXX[y]),
    SumXX = List.Sum(AddedXX[xx]),
    SumXY = List.Sum(AddedXX[xy]),
    Slope = ((CountX * SumXY) - (SumX * SumY)) / ((CountX * SumXX) - (SumX * SumX)),
    Intercept = (SumY - (Slope * SumX)) / CountX,
    PredictedY = Table.AddColumn(AddedXX, "PredictedY", each Slope * [x] + Intercept),
    ErrorSquared = Table.AddColumn(PredictedY, "ErrorSquared", each Number.Power(([PredictedY]-[y]),2)),
    DistanceSquared = Table.AddColumn(ErrorSquared, "DistanceSquared", each Number.Power(([y] - List.Average(AddedXX[y])),2)),
    SumError = List.Sum(DistanceSquared[ErrorSquared]),
    SumDistance = List.Sum(DistanceSquared[DistanceSquared]),
    RSquared = 1 - (SumError / SumDistance)
in
    [Slope = Slope, Intercept = Intercept, RSquared = RSquared]