When I first started working in data, I used to think it was all about the algorithms and tools. Now banging my head against the wall has helped temper some of that. I now see the Data Science role as including at least some management consultancy.
Recently Max Shron in his “Thinking with Data” book, and Hadley Wickham in his Data Science: how is it different to Statistics article, have been raising the importance of training data scientists – and if you are a data scientist, the importance of training yourself in asking the right questions.
These skills are rarely cultivated by a good statistics or Mathematics education. So I have a few methods to help you develop these “soft” skills:
- Business Skills Training: Some universities have these in their programs, and my experience has been that they are a mixed bag.These also help you ask the right questions and practice the stakeholder management and customer facing skills that you will need whether you are a data scientist at Sony, PwC or even a small startup.
- Consultancy clubs: Some universities have these too – generally where you go through some training on leadership or you discuss business cases. These events include the elevator pitch training, networking, communication skills training, training for case interviews. A good example is McGill Consulting club
- A business case book: A few years ago I would have laughed at this as good advice for Data Scientists but now I think reading one of these books. Ideally through discussion with others – is extremely useful for developing the ‘consulting’ part of the data science skillset. I think that it is really important to be able to talk to senior management in their own language – which is often not statistics. And this means understanding a range of things including their own ‘MBA-speak’ , their strategic objectives, their hopes and fears and how they build a mental model of their companies. It also means ‘here is the R^2 values’ is probably not the best answer. Communicating your results is hard and I have found practicing business cases is really beneficial for this.One famous book is the following: “Case in Point: Marc P. Cosentino”. I personally found this book to be a good compendium of examples giving you some of the techniques and frameworks that are taught in business schools or in consulting use. Having a few of these in your arsenal as a Data Scientist is useful for scoping the data science project.
- Finance for non-Finance types: I recently picked up the excellent book “How to read a balance sheet” and there are other examples out there and any good accountancy introduction or online course would be a good training. The reason I mention this is that there is an unfortunate stereotype in business against technical people – and accountancy is the language of business. To succeed in business you need to understand the language. There are other blogs and online courses that you can follow on Coursera also.
- Marketing: Depending on which team you are working with you’ll have to learn some of their specialist vocabulary such as understanding what metrics the marketing team use and understanding how these are calculated.
- Business development: Another aspect that data scientists at fast growing companies need to know is how to integrate data-driven with the growth and development of the company. For these topics I wholeheartedly recommend the “Lean Startup” books. You could start with the “Lean Startup” book itself!
The importance of developing these ‘soft’ skills cannot be underestimated. It takes some time but I hope that my suggestions are a good start for you on your journey to becoming a better data scientist.
Luckily an academic background develops some of these skills – it is impossible to survive in graduate school without collaborating, giving presentations, and grasping other peoples domain-specific vocabulary.
Paedar Coyle is a data analyst based in Luxembourg. His trajectory has evolved across physics, mathematics, and computer science in order to extract value from data, innovate and drive change. His work involves statistical modelling, building Data Analytics Proof of Concepts and working with the software teams to deliver them. He is also involved in the business strategy around data analytics and data auditing.
(Image credit: Hillman54, via Flickr)
