Averaged Shifted Histogram
David W. Scott · Wiley StatsRef: Statistics Reference Online · 2019
Abstract Modern data science employs many advanced algorithms, but always begins with an exploratory data analysis phase. The data summaries typically used in EDA involve point and frequency diagrams, such as a scatter diagram, a box‐and‐whiskers plot, and a stem‐and‐leaf plot. When massive datasets are encountered, alternatives that produce interpretable figures are desirable. In this article, we focus on the histogram, which is an alternative to the stem‐and‐leaf plot. The histogram is useful not only for identifying “structure” in each variable, but also can be helpful in identifying useful transformations, and even for identifying outliers or subsets. However, the quality of a histogram may be dramatically improved by “averaging shifted histograms,” ergo, the topic of this contribution. Furthermore, the EDA phase can extend to more than one variable. In fact, the ASH can be applied in several dimensions in a straightforward manner. We demonstrate how data understanding can be improved for in the multivariate setting.