“The only thing that is constant is change.” That’s from Greek philosopher Heraclitus about
2500 years ago (yes, I had to look up who said it). Data Scientists often use this quote as a reminder that what allowed them to provide business value yesterday won’t necessarily help them make their businesses and customers more profitable or drive new revenue tomorrow. New technologies, new skills and new approaches become commoditized over time and once everyone is doing it it’s the standard instead of the differentiator. As within all business and technology, a Data Scientist must innovate and evolve to continue to drive value.
It’s been a long journey and many of the concepts you hear about today sound familiar. Sure, we are way beyond punch cards and green bar reports, but deriving insight from data is at the core whether we’re talking Management Reporting from forty years ago, Business Intelligence from fifteen years ago or Advanced Analytics today. It’s a continuous evolution that folds in advances in technology with new approaches to provide better/faster/richer insight from information. Data Science is all the buzz today. It includes big data, unstructured data, predictive analytics, machine learning, text/voice/image analytics, Internet of Things (IoT) data sources and real-time data acquisition.
“data scientist” vs “Data Scientist”
One of the key roles to emerge in information systems is the Data Scientist. This is a somewhat controversial term. It was first used by Paul Naur to describe what is more commonly today referred to as computer science in Concise Survey of Computer Methods in 1974. A few years ago Gil Press argued in an article for Forbes that “data scientist” has become merely a buzzword that can mean different things depending on context. Data Scientist is starting to surface in a lot of titles and job descriptions of people who do similar things but with very different backgrounds and intensities.
University programs are springing up that focus on blending applied mathematics with business and computer science, producing degrees that are essentially a cross between an MBA and MS in Information Systems. In fact 92% of “Data Scientists” (note the formal capitalization) have a Master’s Degree or PhD. They bring a grounding in statistical methods and data technologies to the table along with a strong understanding of business processes and drivers to provide insight that is difficult or unavailable with more basic data analytic methods.
Of course there are those without the formal training and deep applied mathematics backgrounds that are also being referred to as “data scientist” (note the use of lower case). These are people who would have more likely to have been called “data analysts” or “business intelligence developers” just a short time ago. But as Gil Press points out, the phrase data scientist is fashionable so it’s appearing in titles and job posting for people not classically trained in the foundations of Data Science.
This is not to say “data scientists” are not important and valuable. They can obtain training from sources like Microsoft and Coursera. Numerous Universities are offering courses and certificate programs that give a grounding in data science concepts and teach languages like R and Python, probability calculations, machine learning models. Such training allows “data scientists” to apply themselves against the same problems “Data Scientist” tackle. Generally a “data scientist” has a little stronger background in data wrangling but might struggle to understand something like binomial probability distributions and more complex mathematical algorithms that drive advanced models because of the lack of deep grounding in the statistics and other advanced concepts that Data Scientists typically have.
As you read through the rest of the post, think of any Pros or Cons that were missed. Be sure and leave a comment and let us know!
Train, Hire or Partner?
Companies seeking to benefit from what Data Science can offer will need to make decisions about how to acquire skills.
Train ambitious, existing staff that are interested in Data Science and build your own “data scientists.”
- Pro – Retain existing talent that already knows the business
- Con – Investment of time and money to get the training and will still fall short of true Data Scientist depth without significant training
Hire “Data Scientists” with formal training from a degree program.
- Pro – Pool for Data Scientists is growing to meet expected demand and you’ll get deep skills
- Con – Hard to find deep specific industry experience so expect time to ramp up and learn specific industry
Partner with a company with deep data roots in business intelligence and data warehousing across broad range of industries that also has strong Data Science skills. Such companies can help identify and solve problems around lowering operational costs, business process optimization, customer acquisition and retention, to name a few. Partners can tackle low hanging fruit to quickly drive value and prove that investment in Data Science is worthwhile. This can allow companies to experience benefits before embarking on hiring and training their own resources.
- Pro – Proven successes with knowledge of various industries and can provide immediate value
- Con – Delays building internal capabilities
If you’re not quite ready to Train or Hire and want to use our Data Scientists let us know. Mariner is ready to help you with that next project. You ready?
So what do YOU think? Feel free to leave a comment with any additional Pros or Cons.