Skip to content

Matt-Stammers/Python_For_Data_Science_Intro

Repository files navigation

Python_For_Data_Science_Intro

Pythonic Data Science Intro For Clinicial Developers

Guys, I'm not going to lie, Data-Science is not easy to learn. Frontend web development is definitely easier, and even getting really good at the backend with something like Python and Django has a much less steep learning curve.

I have been at it for over a year now and I can now fairly reliably deploy basic machine learning algorithms and am gradually moving from python data analyst to data scientist.

The first stage is to learn Python or R, so if you haven't done that I suggest you start by learning one of those languages. I started learning Python on www.codeacademy.com and www.sololearn.com but quickly transitioned to learning it on www.datacamp.com after I did a very expensive bootcamp and came out of it feeling like I knew nothing.at all. With hindsight I now know that this over £400 bootcamp experience was not necessary so I suggest you start by going to Udemy/Udacity and taking a basic python or R course there first.

Regaring which language to start with I recommend Python because it has a much broader range of applications, however, if you are planning to work in a very academic setting or perhaps corporate finance, then R is probably superior at the start (Note however, that it has been my experience that learning R after having a solid grasp of Python was pretty easy (in fact learning all these other languages was pretty easy after getting really good at Python), however, I'm not sure the reverse transition is as simple because Python features both functional and object oriented programming, as well as an enormous array of different applications each with their own quirks. In effect, learning R first in my opinion does not prepare you as well for the 'computer science' side of programming and data science, which please note, is the side that is growing the most as a result of the exponential growth of reinforcement learning and 'AI' which is likely going to become the 'data-science' of the future. I hope that helps you to decide where to start out.

Then it gets a little tricky because you have lots of options once you get going with one of these languages. I waited till the black Friday sales and bought a years subscription to www.datacamp.com which I have to say really helped me to get to where I am (yes, it's not cheap, but think of it like a Gym membership). However, there are also some really great courses on Udemy/Udacity etc which are much cheaper so that is an option as well. I can really thoroughly recommend the A-Z data science course by Kirill Eremenko who I consider to be a really good teacher, and he is very enthusiastic as well.

You then need real life projects to apply your skills towards, as you will see from some of my code. If you don't get your hands dirty you won't progress as quickly. Like with any skill it is the repetitive use that really makes it grow and stick. You need to first fall in love with data and what it can do for people's lives, start small mixing it in with Excel doing data preparations, manipulations, visualisations and 'mining' for insights. Then move towards learning the statistical methods and maths as this is key to transitioning from analyst to scientist. Then learn SQL and database manipulation skills. Then start practicing the machine learning models (you might well want to do a second course on this). Then when this is solid learn about deep learning and network analysis. Note - deep learning is an enormous topic and will take time to master - I have still barely scratched the surface.

That is enough in the present day to call yourself a data scientist (I still don't call myself that because I have too much respect for the term). If you are learning alongside work as a 'hobby' then expect it to take you several years to really master the craft. If you are out of work then go at it and you can probably get pretty good at it within a year or two.

It's a hard skill, but for the determined with a decent mind it can and should be learned as it complements pretty much any job in any industry and in 10-20 years data-illiteracy is likely to be viewed as a very bad thing in the workplace, so getting started now is a very good idea ;)

As always, feel free to msg me on facebook or join the Clin_Dev's email list and send me an email if you have any questions.

Matt

About

Data Science In Python Code Repository For Clinician Developers

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published