What is the Data Science Shop?
The Data Science Shop explains the practice of Data Science, what it can do for your business, how it can operate in a company, and what it needs to succeed.
Data Science Shop
The Data Science Shop details the who, what, where and how of Data Science:
- who – the technical people – beyond Data Scientists – that build Data Products
- what – the classes of Data Products that the Data Science Shop can produce
- where – the technology environments that the Data Science Shop needs to build Data Products
- how – the process that guarantees that the Data Science Shop builds effective Data Products
The Data Science Shop adapts to any company because it can be tailored to any company’s needs, structure, or strategic objectives.
The Data Science Shop provides the structure that aligns businesspeople and Data Science practitioners because it provides a clear understanding of each other’s role.
The Data Science Shop is the product of two decades of organic evolution of Data Science across industries outside Big Tech.
The Data Science Shop relies on 5 roles that use 3 technology environments to build 4 classes of Data Products through an iterative process that involves users and technical teams.
Who
Data Science Shop Crew
A Data Science Shop lodges five distinct roles to build Data Products:
- Data Scientists deliver the advanced science that powers Data Products
- Data Analysts extract insights from Data Products
- Data Engineers ensure that data flows to and from Data Products
- Machine Learning Engineers ensure that optimized computing is available on Data Products
- Project Managers coordinate the process that builds Data Products
Where
Tech Environments
A Data Science Shop uses three technology environments to build, host and maintain Data Products:
- Data Architecture organizes the technology to store, transform and distribute data in a company
- Computing Architecture organizes the technology that provides computing power to Data Products
- Solutions Architecture organizes the technology that makes Data Products accessible to users
What
Classes of Data Products
A Data Science Shop builds and maintains four classes of Data Products, each solving a different kind of need:
- Dashboards automate the display of data summaries or visualizations
- Aswers use advanced science to provide explanations, scenarios, projections or causes
- Deployments deliver automated outputs from an algorithm funneled to some business process
- Custom-built tools that are proprietary applications developed to fulfill a specific business objective
How
Data Product Cycle
A Data Science Shop builds Data Products using an iterative process that involves technical teams and final users in four stages:
- A diagnosis stage that defines the problem to solve and evaluates available data
- A development / implementation stage that builds the Data Product as a solution to a problem
- An adoption stage that ensures the Data Product is embraced
- An evolution stage that triggers a new iteration of the cycle when it identifies additional needs that can be solved with the Data Product
The cycle continues iterating through a Data Product’s lifetime
