How to Use Google (Not the Search Feature) to Forge Machine Learning Knowledge and Keep It Sharp
Before enrolling in the Flatiron School’s Data Science course, I believed that after the 10 months of training and developing projects, I would be a master of everything data science. A level 99 data scientist some may say. As each project was completed, I felt more and more confident about my skills with data mining, cleaning, exploring, feature engineering, predictive modeling and visualization. In my head, each project was like a boss that I had to beat and with each “victory” a boost in experience. However, as the rush of getting stuff done by a deadline waned after graduation, I realized one day after looking at my GitHub contributions that it had been sometime since I completed my last project. For the sake of good ol’ times, I explored my projects to reminisce with the work I had completed. Memories popped up here and there but this time it felt different going through the projects: that confidence had also waned with the time. And unlike your game character whose skills still remain after beating the game, mine were down leveling. Although it is true I learned a lot in the program, the major take away from the program that occurred to me at this moment is that data science knowledge is a skill that will disappear if you stop practicing.
Fortunately, I ran across courses by Google that are giving me a much needed refresher; although there aren’t practice exercises for each section, its theory drive approach with quizzes and videos is just what I need to keep my knowledge skill high. Still, the ones that do have exercises use Google Colab. If you are not sure what that is it is Google’s version of a Jupyter Notebook where you can code in Python and import libraries just as you are accustomed to. So what is covered in the courses? If you are a beginner, I highly suggest you start with Introduction to Machine Learning Problem Framing. In the introduction course, we learn the following:
- Define common ML terms
- Describe examples of products that use ML and general methods of ML problem-solving used in each
- Identify whether to solve a problem with ML
- Compare and contrast ML to other programming methods
- Apply hypothesis testing and the scientific method to ML problems
- Have conversations about ML problem-solving methods
If you are already familiar with some ML or you are not sure where to start just answer one question and Google will direct you to where you need to go.
The Machine Learning Crash Course will take you through ML concepts such as generalization, splitting your data into testing and training sets, classification, neural networks and more! Once you are done with the crash course, you can move on to Data Prep, Clustering, Recommendation, Testing and Debugging, and finally Generative Adversarial Networks. Just when you think that is all, there are examples of how Google uses machine learning in its products, best practices guides, and even more educational material at Google AI. In short, there is a lot you can formally learn from Google than just through the search feature, so check it out and stop your leveling down and continue on your journey to (or keep your) level 99 data scientist!
