Data Scientists and App Development
Six years ago Harvard Business Review called “Data Scientist: The Sexiest Job of the 21st Century”, and we’re now seeing why. Mobile and web apps are increasingly adding features powered by machine learning and natural language processing. But unlike the migration from websites to mobile apps, cognitive requires more than simply retooling web developers into mobile developers. Cognitive requires integrating a new type of team member into application development teams: data scientists. These new colleagues bring a whole suite of tools, libraries, and languages to the app development lifecycle.
For example see this marketing image above which was taken from IBM’s Watson Data Platform landing page. It shows IBM’s recognition that cognitive apps require these four stakeholders to work together. Data scientists are directed to the Data Science Experience while application developers are sent to Bluemix and the Watson Developer Cloud. Business analysts are directed to Watson Analytics. Data engineers are directed to Bluemix Data Connect and are a topic for a later discussion as they are also critical to at-scale application development. The Watson Data Platform combines cloud-based tools specific to each team member so they can interact and coordinate their efforts.
Curiously a quick search for similarly integrated solutions from Google, Amazon, and Microsoft came up empty. Google’s cloud platform for data scientists provides isolated support for building models while Microsoft emphasizes a similarly granular solution with its Azure ML platform that emphasizes integration with Windows. Amazon has the least advanced data science offering and appears more a “me too” IAAS offering than the integrated next-generation solution envisioned by IBM.
So what’s in the Data Science Experience and how does it relate to application development? Click through that link for most details, but here’s a quick list of tools used by data scientists and application developer. The goal of IBM’s Data Science Experience is to integrate many of those tools to simply a data scientists lives much as Bluemix is meant to make development easier for application developers.
You’ll see that common tools and programming language between these two specialties are highlighted in blue. And as you can quickly verity, nothing is blue because there’s no overlap between these two specialties. OK…yes…we could include python as a common language but fewer and fewer application developers are using python for web development. Instead the dissimilarity goes much deeper once we note that data scientists need to understand statistics and linear algebra at a moderately advanced level to be successful. Skills that are not useful for application development.
But if you’re intent on learning data science then don’t hesitate. Multiple online and offline courses provide aspiring data scientists the tools to up-skill quickly. Coursera offers a 6 month Specialization in Machine Learning with a focus on practical use cases while Galvanize delivers a 3 month crash course in data science as well as a 12 month Masters in Data Science program on-site at its many campuses. And if you really want street cred then check out this online course at UC Berkeley’s Master of Information and Data Science.
