The great thing about learning anything technology related is that the internet is overrun with resources. The downside of too many options is the paradox of choice. After scratching at the surface of researching A.I., I decided to get smart and make sure that I strategically step through the different requirements of becoming a data scientist. I finished by bachelors degree 10 years ago, and while pursuing a Master’s Degree seems attractive, the headache of applying for competitive programs and then balancing structured course work and working full time just doesn’t interest me right now.
I’ve come up with my own road map to data science, and wanted to document my plan, and the execution of that plan here – updating and tweaking it along the way. Here it is:
- Brush up on my Python programming. I know enough Python to copy/paste, debug and hack my way through. Machine learning seems complicated enough – struggling with the coding behind it will make it too hard. So I’m going to make sure I’m super comfortable with the language
- Math refresher – Linear Algebra, Calculus, Probability and Stats. These are the fundamentals laid out in the very useful video here. I’ll check out the MIT OpenCourseWare and most likely follow along with the corresponding courses there.
- Udemy Data Science A-Z – Next I think I’ll need a solid understanding of what data science is, and some practical uses of it. That’s why I bought this course on sale, which comes highly recommended by the community.
- YouTube tutorials – Now it will be time to put my knowledge to work. I’ve watched a number of videos by Siraj Raval and his YouTube Channel. He has a lot of practical examples that I’d like to not only implement, but build off of.
- Lastly, I will foray into playing with Google’s DialogFlow to build some real world tools that I can launch to Google Assistant and other services online. This will let me build a conversational AI without needed to program the back and forth dialog and focus my machine learning skills toward analyzing and processing data.