In order to venture out into Data Science with a job/consulting/freelancing/startup I decided to follow the below approach :
- Study Statistics - for Subject Knowledge
- Finish DQ’s “DS in Python” Path - for Python Data Science Knowledge
- Then build any projects - to showcase DS my knowledge
But many are suggesting that this approach takes a long time & that we need to ditch such an approach & dive in right away to do some POCs (Proof of Concepts) & do Kaggle projects. This makes me feel vulnerable thinking that I’ve been & am wasting all my time following my approach.
But simultaneously here is how my thought process clashing with theirs & making me doubtful of my process as well as on the best approach to follow further :
“Don’t we need the knowledge of Subject & Python to be able to dive into POCs & Kaggle Projects?? How could one dive into them right away without such knowledge?? Am I wasting a lot of time with the approach am following??”
Need your expert help/guidance on the topic guys!! Please help! Feeling confused & overwhelming due to this.
Few notes from my experience as a person who went the same way including totally similar thinking.
- You can’t learn everything. So learn the very basics for the beginning and the rest you will learn on the way.
- Start DOING something (vs LEARNING everything) as soon as possible. I’ve spent couple of years (since I had such an opportunity) trying to learn as much as I can… And the very first real project I tried to approach clearly showed that I do not remember a … well, most of what I was learning It’s only when I attempted to do something and I learnt something while DOING - that things I do remember now. In other words - practice is the best teacher.
- So try many various small projects to find what you like doing most, as DS is very wide field. Then select this one specialization and look where you can do something useful. And do this.
- And do not give up! It’s in our hands to change everything.
Thank you so much for the response ‘ranklord’. Resquest you to respond on the below as well - which could give me more clarity :
Can I assume that the below activities would come under your 1st advice (learning the basics) & are not a waste of time?? :
- Trying to finish the DQ’s “DS in Python” Path (including the Guided Projects)
- Completing the book “Statistics without Tears”
- Completing the book “Probability without Tears”
- Completing the book “Hands on machine learning with scikit-learn and tensorflow”
You are welcome!
- Do finish the path! This will give you solid base to work on projects and understanding of what to learn further. I have finished Coursera’s DS path because I didn’t know about DataQuest at that time There is also Udacity with Nanodegree programs, but it’s quite advanced and expensive. DataQuest is right quality, right mix of theory and practice and right price! This opinion is also based on my experience of using multiple platforms. See my bookmarks list only for “Learning” section (if you click on image, it will show you more :))
- And 3. These would be useful addition to your knowledge and understanding and definitely not a waste of time. I am not familiar with books you mention and may extend your list with those, also good ones:
- “Elements of Statistical Learning” by Trevor Hastie, Robert Tibshirai and Jerome Friedman
- Three free books from here: https://www.openintro.org. You can do them all or just any
- “Statistics for Data Scientists” by Peter and Andrew Bruces
- “Statistics in a Nutshell” by Sarah Boslaugh and Paul Andrew Watters
- Not familiar with this one either
And once again, please, do not wait until you finish any course and / or book or all together! There is NO such state as “I have learnt everything and now I know” There always will be something you do not know yet.
Learn and do small things, grow your GitHub, apply what you’ve learnt to your projects, get stack and learn again Also, your potential customers / employers will want to look at your projects done, and not on your certificates
This road is tough and so amazing!