This week, we continue with the fourth issue in our ongoing data preparation series by addressing two important anomalies commonly found in sample data: (1) concentration of values – either a tight range (e.g. proportions) or a wide dispersal over several order of magnitude (e.g. population, income, rainfall volume, etc.) and (2) the mean/variance relationship.
For this tutorial, we’ll start by going through a few different value concentration cases, explain their impact on analysis and forecast, and, finally, present common transformation methods and discuss concerns for mapping results from a transformed data scale back to raw data.
You can find more details and a step-by-step tutorial, along with a downloadable spreadsheet and PDF, at the link below:
