Imagine you have some data. Maybe you’ve done a web redesign and you want to do some A/B testing to see if by redesigning the page your visitors spend more time reading your content. You set up an experiment in which visitors are randomly sent to either the old or new design, and you measure … Read more
Before you report your conclusions about your data, have you checked whether your 'actionable' figures occurred by chance? The Kruskal-Wallis test is a safe way of determining whether samples come from the same population, because it is simple and doesn't rely on a normal distribution in the population. This allows you a measure of confidence that your results are 'significant'. Phil Factor explains how to do it.… Read more
Although linear regressions can get complicated, most jobs involving the plotting of a trendline are easy. Simple Linear Regression is handy for the SQL Programmer in making a prediction of a linear trend and giving a figure for the level probability for the prediction, and what is more, they are easy to do with the aggregation that is built into SQL.… Read more
Statistical calculations in SQL are often perfectly easy to do. SQL was designed to be a natural fit for calculating correlation, regression and variance on large quantities of data. It just isn't always immediately obvious how. In the second of a series of articles, Phil factor shows how calculating a non-parametric correlation via Kendall's Tau or Spearman's Rho can be stress-free.… Read more
Some people will assure you that you can't do any serious statistical calculations in SQL. In the first of a series of articles, Phil factor aims to prove them wrong by explaining how easy it is to calculate Pearson's Product Moment Correlation.… Read more