This week’s tech tip is on interpolation, one kind of data transformation data scientists use to create new data points within the range of a discrete set of existing data points. This technique is often used to transform quarterly data to monthly data. Presented below are three common methods of interpolation that convert the original quarterly data (black) into monthly data (blue). As observed, the transformation adds two points between each of the original points. Where these points lie depend on the method. The method employed may have strong implications on analysis and should be dictated by the application.
Constant: This method assumes that the monthly value is constant for the entire quarter.
Linear: This method linearly interpolates between quarterly points.
Cubic-spline: This method smooths out the shape and removes sharp changes.