The Different Types of Data Scientists
Previously, we established that the definition of the term Data Science requires you to understand the difference between the key skills that constitute the technical practice of analysis within the field. In addition to the hard skills of Predictive Analytics, Machine Learning, and Business Intelligence, the quintessential component of Data Science (DS) is fluency in Interdisciplinary Communication. Humans are naturally curious, (even business people!) and acceptance of a technical concept is more likely to follow understanding than trust.
With this in mind, whether you’re out and about looking to hire DS talent or just trying to manage your existing resources more effectively, it’s important to understand how the core competencies of Data Science interact with each other. By combining different competencies, we can establish archetypes that describe the kind of background effective DS candidates and practitioners may possess.
Meet the Team
In addition to fluency in Interdisciplinary Communication, the most important distinguishing trait of a Data Scientst, a strong DS candidate with high long term potential possesses a strong interest and ability in two of the core competencies. Junior-level data scientists, or practitioners that are just starting out in their careers, may only possess demonstrated ability in one of these fields (and that is fine!), but should possess a strong interest in learning more about the others. Using their position to learn about other key skills allows them to develop into some of the archetypes below.
Only unicorns have a high level of competency in all three of the core skills. Unicorns are out there, but don’t bank on building your practice around one.
The Academic
The Academic is strongest in Machine Learning and Predictive Analytics. They have either a history in academia or a strong and lasting curiosity about learning. The stereotypical Data Scientist job description typically looks for these folks, especially if there are hard requirements around education, such as Ph.Ds in mathematics or computer science. Academics typically have a strong grasp of the mathematics that underlie both Predictive Analytics and Machine Learning, but are less savvy about the soft aspects of business, such as market momentum or consensus building within the organization. They usually have adequate technical programming skills, but less so than other archetypes.
If your organization is in a position to publish in the literature of your field or evangelize about the power of your solutions, your Academic is a candidate to lead that effort. They get why it’s important, what’s required to make that happen, and the standards to which peer-reviewed journals hold themselves.
The Hacker
The Hacker exhibits a high level of skill in Business Intelligence and Machine Learning. Typically, Hackers develop out of the operational practices of businesses. They may be business analysts that took a particular shine to technology, or operational analysts who took it upon themselves to optimize processes above and beyond the call of duty. Crucially, hackers develop themselves – they may not have the deepest understanding of the mathematics underlying Machine Learning or Predictive Analytics, but they know how it works at a technical level, and how to implement it. Hackers tend to have superior programming skills to other archetypes.
Let your Hackers lead the process of defining and bringing your DS infrastructure into production. They’ll be interested in the development of the system, and can play a role in teaching development teams the best ways to implement your DS product.
The Quant
The Quant is particularly strong in Predictive Analytics and Business Intelligence. Quants have a knack for turning mathematical models into useful business insights and defining business problems in such a way that they can be solved using Data Science. Many Quants have come out of the Finance industry seeking a change in environment, after finding their surroundings stifling or uninspiring. They may have MBAs or training in management consulting; indeed, Quants speak the language of business. They tend to have moderate experience in applied statistical languages (such as SAS, R, Stata, or Matlab) and light experience in general purpose programming languages (Java, Python, C, etc…).
Quants are very useful in building support and momentum for the DS practice in your organization. Let them scrutinize an existing process and come up with a few ways that it might be improved using data. To this end, the typical background in Finance or Management Consulting will come in handy.
On Continuums
None of this structure is to suggest that all Data Scientists will fall neatly into these categories. Being that one of the more important traits of a Data Scientist is curiosity, your candidates will typically have an abiding interest in all three fields to some degree. Instead, these archetypes are more useful in assessing where you might have a gap in your team.
Are you having a tough time achieving buy-in from other stakeholders in the business that your work is important and valuable? You may be in need of a Quant.
Is your DS structure inefficient and holding up the implementation of your innovations in production? Search out for a Hacker.
Are you having trouble convincing others in your industry that you’re onto something special? Look for an Academic.
If you haven’t already, read some of the other pieces that I have written on the topic of managing data scientists, or get in touch if you’re interested in exchanging some ideas.