Before I mention any tips or suggestions for how to more effectively teach data science, I want to start by saying that teaching data science is not an easy task. With so many application domains and encompassing subdomains, no one class or series of classes can ever hope to cover the world of data science. Additionally, I want to highlight that everything here is simply my view of the data science world — a world I have studied, but only recently been emersed in.
With that being said, I still believe there are a number of useful suggestions that can help supplement student learning in the classroom when it comes to data science. As I transition from classroom learning to the business world, I can’t help but notice the disconnect between the two and how this gap could be mitigated. Below are 10 tips that I believe will not only further enhance the data science learning experience, but also help to prepare individuals for careers in data science.
1. Name Game
One of the first things you learn in computer science is that variables should be appropriately and succinctly named. A name too long is inefficient to type and makes code clunky. A name too short may not be representative enough of the variable’s contents. There seems to be some sort of middle ground sweet spot, but how do you find it. My advice for the best naming practice is to create the shortest name possible, such that a person other than yourself (with limited to no knowledge of the code or project) can accurately tell what the variable represents.
While the worst examples of poor naming can cause the developer themselves to forget what a variable represents, most times disobeying the rules of the name game will make things confusing for others. Poor variable assignment effects can also be seen outside the environment of one’s code too. Databases often contain a plethora of examples of poor attribute naming, containing unexplained acronyms, excessive_use_of_underscores, etc. Sometimes these tactics are necessary, as variables can represent rather complex ideas and figures. However, when these practices are utilized, it is important to document these areas of confusion, so that future users or editors of the code/database (including yourself), will not be lost.
2. Commenting Beyond the Code
Variable assignment is only one potential source of confusion when coding or developing a data science project. Due to the fact that certain code blocks can become rather convoluted, commenting is usually seen as a way to explain some of the intricacies. However, oftentimes comments are randomly strewn throughout code here and there and used as short term reminders for the developer, rather than designed for users or other editors of the code later on. Having readable and clean code is arguably just as important as having code that will run smoothly and therefore should be highly stressed to students. The way I see it, the more one explains what is happening within their code, the longer they can expect that code to be used in practice. If only you yourself can understand how it works then it is more likely to die once you commence working on it.
#data-science #teaching #learning #python
