fbpx
Practical Ways to Integrate Data Science Into Your Organization Practical Ways to Integrate Data Science Into Your Organization
This ODSC West 2018 non-technical talk “Practical Data Science,” presented by Michael Galvin, Executive Director of Data Science Corporate Training for Metis, discusses the... Practical Ways to Integrate Data Science Into Your Organization

This ODSC West 2018 non-technical talk “Practical Data Science,” presented by Michael Galvin, Executive Director of Data Science Corporate Training for Metis, discusses the practical steps that need to be taken to successfully integrate data science into an organization, as well as some struggles and pitfalls that commonly occur along the way. You’ll hear real-world experience about building data literacy skills, collaboration, and investments in existing talent, as well as why these are important elements in building a successful data science function.

[Related Article: Most Data-Driven Cultures… Aren’t]

The talk is divided into 4 sections:

  • Hiring and Retaining the Right People
  • Educating Stakeholders: Lifecycle and Expectations
  • Data Literacy
  • Creating a Business Culture that’s More Data-Driven

When discussing the required skillsets for members of data science teams, Galvin provides the usual data and analytics skills including a number of “soft skills” like “grit, problem-solving, adaptability, and passion,” all of which can lead to the core of a successful team. The keyword is “team,” and Galvin carefully navigates how hiring a single “unicorn” is counter-productive to this team mentality. Rather, the list of skills should be spread about the various members of an interdisciplinary group. 

Additionally, Galvin reviews many of the common roles for members of data science teams including: data scientist, data engineer, data analyst, ML engineer, etc. There are few standards for defining these roles, so there’s often confusion and difficulty when trying to fill these roles with individuals who have misaligned skills. A key hire, the data science leader (not necessarily the most technical person), should be in place before the rest of the team.

integrate data science

From “Practical Data Science,” presented by Michael Galvin

The presentation then moves on to discuss the importance of educating project stakeholders on the data and analytics lifecycle (which I like to call the “Data Science Process”) including such areas as: problem statement, data collection, exploration, modeling, and deployment. What you need to avoid is when a stakeholder says “Here is a data set, now go create cost savings, profits, or a product out of it.” This lack of a firm sense of the problem to be solved usually never works out. 

You’ll also learn from the presentation how important it is to be clear that the lifecycle is a continuous one, where models need to be managed and monitored, and models must be retrained over time using newly collected data. 

The following slide from the presentation stresses the importance of establishing data literacy across the organization. Shown is the gap between data literacy skills vs. demand. The presentation covers methods for closing this gap.

integrate data science

From “Practical Data Science,” presented by Michael Galvin

Once data literacy is in place in the organization, Galvin discusses how it’s important to produce a culture centered on enterprise data assets, specifically how you use data to drive decisions throughout the organization. The presentation includes a number of well-define methods for building this culture.  

To take a deeper dive into perspectives for building data science teams along with their capabilities within organizations, check out Michael Galvin’s compelling talk from ODSC West 2018.

[Related Article: Why You Should Be Using Sentiment Analysis for Social Media and Decision Making]

Key Takeaways:

  • Hiring and retaining the right people are both instrumental to a successful data science effort 
  • Making sure project stakeholders are well-versed in the data science process, and also providing clarity on how expectations for success are managed properly 
  • Understanding the importance of data literacy across the organization
  • A checklist of methods for establishing a culture centered data 
Daniel Gutierrez, ODSC

Daniel Gutierrez, ODSC

Daniel D. Gutierrez is a practicing data scientist who’s been working with data long before the field came in vogue. As a technology journalist, he enjoys keeping a pulse on this fast-paced industry. Daniel is also an educator having taught data science, machine learning and R classes at the university level. He has authored four computer industry books on database and data science technology, including his most recent title, “Machine Learning and Data Science: An Introduction to Statistical Learning Methods with R.” Daniel holds a BS in Mathematics and Computer Science from UCLA.

1