How I am learning Data Science - some learning resources and methods
I have been focusing on learning DS for 5 weeks which is preparing me for ultimately landing a job in this area. After referenced many posts on how to be a DS like this post on Kaggle. I have used most of the time learned Python, statistics, and used minimum energy on Git, AWS, docker as support tools for DS. There are more than enough free or cheap courses online, I hope the following review information about resources I have used could help a bit for anyone with DS.
From what I have seen, all of Kaggle's courses are the first must-have for anyone who wants to be a DS(maybe a bit later for Computer Vision and NLP).
Then I use a lot of Datacamp.com. It is an online course platform major in Python, R, SQL over 3 hundred courses. The biggest pro is the price is cheap. When on sale (It is on now to 12 of Jan), A premium subscription costs USD 149 if your IP falls in the US and Canada, or $99 if you in other countries. The courses cover everything in DS from basic syntax in SQL, Python, and R, to data cleaning, data engineering, ML, analysis methods, and some projects to follow, only missing a systematic statistics theory and concepts course. It uses a lot of Fill the Blanks in coding, which is good for fresh beginners. However, it is like to learn with the driver's license coach, you don't need to know how to fuel, maintain, read a map, and plan the route. you probably won't even remember where you have just driven. So you better kick your coach off the car and drive what you just drove and maybe how to find a parking place, how to fuel and replace the spare tire. So I study most of the courses twice, in the second time, I would clean every line of code and rewrite all the code again and again until I can write the whole piece by myself.
Another best practice in learning is to list what I have learned like a knowledge tree. There are many benefits. It helps me to realize the relationship of what have learned, it links unstructured points to a knowledge net. Also, it makes reflash and use what I have learned conveniently. Last, it helps me to check what skills I haven't touched.
Machine Learning A-Z™: Hands-On Python & R In Data Science and Machine Learning Practical: 6 Real-World Applications are highly recommended ML for DS by many. I will join these a bit later.
My SQL skills are relative enough for being a DS. Now, I occasionally learn some SQL courses in DataCamp, such as Cleaning Data in PostgreSQL Databases and how to optimize queries. I actually highly recommend a free course in Udacity, SQL for Data Analysis for everyone who wants to work in any data analysis area.
As for statistics, I recommend reading these 3 articles in Medium to have a quick understanding of Statistics for DS. ML+ also have serval independent articles help me to clarify some important concept in Statistics, such as p-value and T-test. And if you feel that need to learn from easy and basic, <head first statistics> is for you.
As for project practices, I found this on Medium, a step by step recording for a financial investment using docker project, AWS and API. If the author could finish all 4 parts, I believe hand-on all the steps would provide me a rounded and complete practice in DS and is big enough (as an individual learning DS) that could write on my resume.
I believe focus and minimization are very important for success. So the above areas are everything I am working on and plan to 'finish' them in another 100 days. Then I will pick up and dig a bit deep in Tableau, PBI, SSDT, and starts SAS or R. Of course, there are many important areas needed to be job-ready besides these tech skills, including job-hurting and soft skills. So I will try to talk to some Data Scientists to understand what they are doing in their work and listen to their experience and views about DS.
I know DS is not as hot as a couple of years ago, which means very difficult to chase and jump on the bus. But if you never try, you'll never know, there is nothing to lose. I welcome any suggestions, advice, and opinion to me. Thanks in advance.
