“Building a data science product is quite like constructing your home. Using this analogy let’s look at the five roles and skills that the best data science teams hire for.”
The CEO of a large financial services firm was a big supporter of advanced analytics. He decided to get his organization started on the path towards data science.
How did the organization plan this journey? By recruiting data scientists, of course! They hired 1000 data scientists, each at an average cost of $250,000 a year. Data scientists are hard to come by, so the CEO was very proud of this achievement.
Several months and millions of dollars later, the business benefits were not there. And upon further investigation, the organization found that these ‘experts’ were not data scientists at all!
McKinsey reports that neither this firm’s CEO nor the human resources group understood the data scientist role. They naively assumed that a bunch of high-priced technical experts could transform the organization into a data-driven one on their own.
Data scientists often come with excellent machine learning skills, but they stumble at choosing the right business problems to solve. They struggle to scale the algorithms in production. And, they fail terribly in translating the data insights into a format that business users can consume.
Data science is a team sport. Every team must hire five roles if they are serious about data-driven decision making.
Building Successful Data Science Teams
Creating a data science solution is not very different from building your home. It’s intuitively that simple to understand but practically as tedious to execute. We will look at the five data science team roles by contrasting them to the five roles needed to construct your home.
1. Data Translator – The “Architect”
An architect is one of the most critical roles in construction. She discovers the aspirations of homeowners and determines the feasibility of land use. Translating user needs into building sketches, she lays the foundation that the rest of the construction team can build upon. She ensures that the home is functional, safe, sustainable, and delivers on the promise.
Like an architect, a data translator is the best hope for a business in protecting their investment in data science. The data translator understands a user’s business needs and helps identify the most relevant projects to execute. She translates the requirements into a format that the data science team can understand. Her role continues throughout the project and is crucial in creating an actionable end-product that users can adopt for decision making.
Skills Needed: Data translators are domain experts who are proficient in business analysis. With a strong understanding of data, they are excellent team leaders and communicators. They are skilled in general-purpose data tools such as Microsoft Excel.
2. Data Scientist – The “Building Services Engineer”
A building services engineer designs and creates internal systems that make buildings functional and efficient. With responsibility for systems such as HVAC, water, power, and control systems, they build the heart of a home. Today, these experts are the brains behind intelligent systems that power self-regulating homes.
Similarly, a data scientist designs and creates the heart of a data science application. With responsibility for producing business-relevant, actionable insights, he harnesses the power of data analytics. He uses various statistics and machine learning techniques to embed intelligence and continuous learning ability into solutions.
Skills Needed: Data scientists are proficient in exploratory data analysis, statistics, machine learning, and AI techniques. They often know tools such as R, Python.
3. Information Designer – The “Interior Designer”
An interior designer works with the architect and engineers to create a functional and aesthetically pleasing interior. She details what the space will be used for and draws up rough plans. By iterating on them, she develops detailed designs and identifies the kind of building materials to use.
An information designer makes the data science solution functional and pleasing to use. Starting with the information architecture, she develops mockups and detailed design prototypes. She makes the data insights consumable by identifying the right kind of charts, interactivity, and visual design to use. She is a master storyteller with data.
Skills Needed: These experts in information design are skilled in aspects of interaction and visual design. They use design tools such as Sketch, Adobe Illustrator or exploratory visualization tools such as PowerBI.
4. Machine Learning Engineer – The “Civil Engineer”
A Civil Engineer brings the design skA Civil Engineer brings the design sketches to life. He inspects and evaluates all designs to ensure that they are implemented in spirit. He documents processes, maintains construction logs, and ensures compliance with industry standards.
A Machine Learning (ML) Engineer builds out the working application. At the backend, he connects to the data sources, packages the machine learning modules, and integrates with all other systems. He brings the front-end to life through a functional and efficient user interface. He documents the code, maintains logs, and adopts software engineering standards.
Skills needed: ML Engineers are DevOps experts with strong backend/frontend coding skills. They use languages such as Python, JavaScript, SQL, and are skilled in using cloud platforms such as Google Cloud.
5. Data Science Manager – The “Construction Manager”
A Construction Manager oversees the project and keeps all commitments made to the homeowners. She owns schedules, maintains quality, and manages finances. Her job is to ensure that all roles not only deliver their responsibilities but also collaborate well. She handles workplace issues, maintains morale, and ensures workplace safety.
Similarly, a Data Science Manager is the shepherd of a data science team bringing all the roles together and empowering them to give their best. She keeps all client commitments and maintains communications. She ensures timely quality deliveries. More importantly, she is responsible for change management and adoption of the solution by business users.
Skills Needed: Data Science Managers are excellent project managers who are skilled in change management. They have a good grasp of business analysis and the approaches to frame data science solutions. They use project management tools such as Microsoft Project.
Building A Data Science Team That Delivers Business Impact
We’ve looked at the roles needed to construct your home. What about the raw materials you need to build, for example, bricks, steel, or wood? In our analogy, data is the raw material that goes into making any data science product. Data must first be collected, curated, transformed, and stored before you can act upon it.
The discipline of data engineering and roles such as data engineer take care of these activities. In this article, we’ve focused on the discipline of data science that helps extract business value once your data becomes available. We’ve seen the areas of expertise that the five roles need. All of them must additionally possess some non-negotiable, fundamental skills.
Every person must have a basic domain orientation to understand the business problems they are solving. They should be data literate with an ability to understand, interpret, and converse in data. Finally, they should all possess excellent communication and presentation skills to collaborate amongst the team and their business users.
Today, there is a lot of buzz around machine learning and artificial intelligence. All this hype has led to disproportionate attention on the sexiest role of the century, the data scientist! However, you need cross-disciplinary skills to make data science work for your business.
Every team needs each of these 5 data science roles to create a useful, consumable, and actionable business solution.
originally posted on forbes.com by Ganes Kesari
Author’s Statement: I’m the Co-founder and Chief Decision Scientist at Gramener. I advise executives on data-driven leadership. I help transform organizations by building data science teams and applying decision intelligence.
