The ultimate, step-by-step guide to learn data science from scratch and prepare for your future job.
It is organised in a step-by-step way. Do not skip ahead. Do all the exercises that are demanded. Whenever possible, write with pen and paper instead of on the computer only. This will make you remember more of the material studied.
In general, prefer official documentation, if it is well-written, over external tutorials. Be very selective about the educational materials you consume. If you learn the wrong stuff, it may be worse than if you did not learn anything at all. It is harder to re-learn the correct behaviour/skill/knowledge. Spend lots of time selecting what you study.
Skim through this repository and its content before deciding to go ahead with it. Finish what you have started.
Do not expect that you will get a job just because you finished this course. Be realistic but also optimistic.
Data Science is a multi-disciplinary field. Instead of going through a standardised curriculum, such as Mathematics or Engineering, you have to be initially analytical in your approach to learn separate things really well and then synthesise them together.
Always balance your study with practice. Make it at least 50% theoretical study and 50% practice.
Examples of practice:
- Web app that makes data science related visualisations
- Open source project that has numerous reputable and active contributors
- Internship or work related project that you can display publicly
Make sure that when you work for employers to discuss the possibility to be able to publish your work as much as possible to the wider public. Not only does it serve the public more, you will also be able to benefit from others seeing your work that you did and for which you get paid. Of course, use common sense to avoid giving out confidential information.
A word of caution: study alone will get you nowhere. Data Science is a skill that you have to practice everyday, much like playing the piano. If you do not practice Data Science, you lose your skills gradually. That is also one reason why experienced and professional data scientists are sought after. It is hard to maintain such a multi-disciplinary skill over a prolonged period of time.
Try to make your study time more about recalling rather than reviewing. Test the key concepts and contents of Data Science using flashcard such as Anki. Test more than you learn initially. Go through the concepts, lectures, readings, books once and make flashcards to test yourself over the next few days. No need to review. Recall.
When you learn coding, do not copy and paste. Do not. Instead, you read the code and re-write them. Yes, it takes more time but you will remember more. Ultimately, you are looking for being able to write simple codes out of memory without reference. You can test this by making yourself offline and write a simple app from scratch. If you have to copy paste code often, it is not a good sign. Programming languages are languages too. If you are fluent in a language, you do not need to refer to dictionaries all the time.
-
Curated AI/ML/DL resources, contributed for the community, by the community
-
Curated Data relatd resources, contributed for the community, by the community
You are welcome to make a PR to this repo. As you contribute more, you can also apply to become an official contributor. Please go to our Facebook Group and tag "John Wu" and ask for the permission to become an official contributor. You can also reach me that way to get tech or usage support. I am available during working hours for direct chats on Facebook Messenger.