The entire platform revolves around the idea that data skills are muscles, most of which can only be strengthened through exercise. All the theory in the world will not write a Python script, and watching yet another video on machine learning will not help someone choose the right class of model. However, practice does help.
Each exercise is designed with multiple levels, just like one might add weight to their exercise machine. And, like a trainer will show you how to use a machine, the platform will explain how to manage each of the exercises.
The Gym is set up so that users can follow a standard path, or skip around for their own personal fitness needs. After all, tennis uses different muscle groups than football, so training regiments will vary accordingly.
The exercises themselves currently involve basic Python programming. At some point we will incorporate a low-code version, but for the beta tests, a little programming is required. We don’t apologize, but rather embrace this, as we expect all our users to achieve advanced data science positions, where they will need Python anyway (but no pressure, and wherever you are with your skills is right where you need to be today – you’re doing great!).
One thing that you won’t find at the gym: instructions. There are a million quality resources out there, and we will probably point you to some of them (or you can check out the forum). But people have been teaching linear regressions for 200 years, and we don’t think we can do better than them. In other words: you don’t learn to walk at the gym. Come back when you’re on your feet, and we’ll get you trained up!
If you’ve ever talked to a data expert, you might have noticed that they have a feel for new data sets. You could put three of them in separate rooms and show them the same novel data, and they will likely ask surprisingly similar questions. “Is this survey data?” “How did you measure that?” “What was your sample size?”
Some of this is driven by knowing the theory, but much comes from experience. Our second major goal is to introduce users to “workout buddies.” Of course, “buddies” are not actual people, but rather real data sets, with exercises to show you their unique characteristics. As learners get to know these buddies better, they will start to see the similarities and differences between data sources. This will, in turn, hone their instincts, further preparing them for the data world.
All of this is still being built, and our staff is rather limited. But please drop us a line if you have questions, suggestions, or want to get involved! jimbodonahue@gmail.com