Guest blog post by Bernard Marr.
In my last post, I explained the difference between what I consider the two core types of data scientist – strategic and operational.
Broadly speaking, they require many of the same skillsets – but the distribution of your expertise and experience within these skillsets will vary, depending on whether your role is more strategic, or operational.
So here’s an overview of what I feel are the five essential skillsets that are required from any data scientist who wants to be competitive in today’s market. Some of them are more valuable to organizations with a need for strategic planning of their data-driven enterprises, and some are more valuable for organizations needing people who are willing to get their hands dirty with the nuts-and-bolts mechanics of data.
However with a broad understanding of all of them, as well as an idea as to where your own particular strengths and interests may lie, you’re in a strong position to sell yourself to the growing number of companies looking to hire top-notch data talent.
Particularly for those whose calling lies towards the “strategic” end of the spectrum, a thorough understanding of what keeps businesses ticking – and more importantly, what causes them to grow – is an essential field of expertise.
You should feel comfortable with the KPIs and metrics that business strategists use to evaluate every aspect of an organization, from its stock performance to its human resources. You also need to be able to evaluate what it is that makes your business thrive and stand out from the competitors – and if it doesn’t, you need to have ideas about how to make it so.
I also include communication skills under this heading, although of course they are important across all disciplines. But particularly in business, the ability to clearly put across the ideas so that every member of the team knows what you are doing, why you are doing it, and how you are going to achieve it, is essential.
The ability to spot patterns, discern the link between cause and effect, and build simulated models which can be warped and woven until they produce the desired results is the domain of the both the operational and strategic data scientist.
Once your distributed storage is threatening to spill over with the reams of structured and unstructured data your machines have pulled in for you, it’s still going to take a human brain to make any sort of sense out of it. As such, you’ll need a thorough understanding of interpreting the reports and visualizations wringed from your reams of data.
You will need a grounding in industry-standard analytics packages such as SAS Analytics, IBM Predictive Analytics and Oracle Data Mining and a firm idea of how to use them to spot the answers to the questions you’re asking.
If an eager, fresh-faced graduate from college or high school has had any exposure to the world of data science before throwing themselves into the workforce, it will probably have been in the computer science lab.
Data is of course essential to everything that computers do, so it’s natural that those with an interest in programming, networking and system architecture often gravitate towards analytics and predictive modeling.
And it’s a good job, too – as techie types are needed for everything from plugging together the cables to creating the sophisticated machine learning and natural language processing algorithms – or whatever happens to be pushing the boundaries of what we can do with the help of our silicon-based assistants today.
In particular, candidates with a firm grasp of key open source technologies – Hadoop, Java, Python etc. - are keenly sought, as these are the foundations of many organizations’ plans to use data to dominate the world.
A statistician’s skills come into play in just about every aspect of an organization’s data operations. They will help to define relevant populations and appropriate sample sizes at the start of a simulation and to report the results at the end. Statistics (and its big brother, maths) is another academic wellspring from which gushes a torrent of talent into the data science workforce.
Whether your role is strategic or operational, a basic grasp of statistics is essential, but if you veer towards the operational, a more thorough education in the subject will be highly desirable.
Mathematics, too, will come in very useful – despite the huge increase in the amount of unstructured and semi-structured data we are analyzing, most of it still comes out as good old-fashioned numbers.
In this sense, creativity is the ability to apply the technical skillsets mentioned above, and use it to produce something of worth (such as an insight), in a way other than following a pre-determined formula.
Anyone can be formulaic – today, businesses want innovation that will set them apart from the pack, both in terms of their corporate results and the image they present to their consumers.
The possibilities made available by the application of data science are constantly evolving. With the explosion in the number of organizations realizing the advantages of leveraging data for insights that will prompt growth, people able to come up with creative methods of applying these skillsets will have a bright future ahead of them.
I hope you found this post useful. I am always keen to hear your views on the topic and invite you to comment with any thoughts you might have.
About : Bernard Marr is a globally recognized expert in strategic metrics and data. He helps companies and executive teams manage, measure, analyze and improve performance.