I’m acutely aware that there are still so many other skills I’d like to build up. The road map below is one that’s probably seen on many data science guides. It can be traced back to Swami Chandrasekaran and it nicely summarises 10 “domains” to becoming a data scientist. I’m pleased that I have at least a dent in all those areas so I’m off to a good start. Given how much there is to cover, it’s clear that no two data scientists could ever be the same as there are plenty different directions each one could branch off. Personally, machine learning feels like the most appropriate step for me just now.
I’ve been casually taking an open online course on machine learning from edX delivered by Professor John W Paisley of Columbia University. A colleague recommended this course over the more commonly referenced Andrew Ng course at Coursera as the mathetical content is deeper. I certainly felt like some parts flashed me back to my maths degree as the presenter walked through some equations in detail. I followed the content and it was absolutely useful in connecting the principles of an idea with the fundamental maths that support it. I actually think the maths makes it easier to remember a particular concept too. For example, decomposing the ridge regression into the least squares form with an additional sum of squares component highlights the behaviour and reason for the introduction of the penalisation term over the usual least squares regression.
Hoping the momentum stays with me all the way until I become a golden unicorn.
Whizzy Data 
