This website teaches data science by building the concepts up from scratch. Here, we build the concepts up from the statistical beginning. This is because I think that the fundementals are often overlooked because it is cooler to jump on the shinyest technique of the day. The problem with the shinyest techniques, however, is that they are often not the most suitable nor the easiest to set up. Often, older techniques are much better but choosing the best requires strong fundamentals. However, we need one tool to get going because doing data science by hand would be no fun.
To make things fun, we rely heavily on the Julia programming language. For me, this language makes things fun for a few reasons. Namely, it is completely free to use, it is extremely powerful and extremely expressive. For example, let's plot a quadratic function from 1 to 100:
With this we're already hitting exactly why data science is so difficult. It cannot be done by hand, so it requires knowledge from statistics, programming and mathematics. And actually, statistics is also largely psychology since it depends on the audience what the best statistic is. This website will try to find a fair balance between these topics. Wherever things do not provide value to the main story, they will be hidden. For example, when talking about distributions, most plotting code will be hidden; only the plotting output will be shown.
Built with Julia 1.7.2 andCairoMakie 0.7.2