The author speaking again: how was your EDA? You found where the drop came from, didn't you? This is great, and we are going to see what the boss thinks about it in a few pages, but let me just summarize here which topic you have been working on with your colleague and what you have learned.

First of all, you were introduced to the concept of EDA and how it can be included within the data analysis process.

You then learned about summary EDA and actually performed it on real data, focusing on quartiles and median, mean, variance, standard deviation, and skewness. For all of them, you first got a sense of what these summary statistics are about and how they work. Finally, you learned what the relevant functions are and how you have to employ them in order to compute each and every one of these statistics.

As a last step within the summary EDA field, you discovered the Anscombe quartet, which is composed of four different datasets sharing a lot of identical summary statistics, even if being very different when actually looking at them through scatterplots. This was shown to you as a way to highlight the limits of summary statistics alone, and to introduce you to the relevance of the graphical EDA.

Moving on to graphical EDA itself, you looked at histograms and boxplots as a way to observe a distribution of univariate values, and particularly as a way to detect outliers. You actually gained a closer look at how the boxplot values are computed and stored.

You finally took a look at multivariate data through scatterplots and linecharts, also looking at advanced ways to customize a plot in order to better convey the desired message and therefore make it ready to be shared with relevant shareholders.

Quite a lot of stuff for a newbie at the Hippalus company! Nevertheless, I have the presentiment that the most thrilling part has still to come: are you ready to discover what caused the Middle East drop?