Data Science Roles and Career Paths: Advice From AT&T's Boykin
In a 2024 AI Connect session, AT&T Director of Advanced Analytics Dr. Frances Boykin explained the roles around data science, from business analysts to data architects, and how newcomers can enter the field. She said you do not need to be a statistician and can start in an analyst role.
By The AI Enterprise Magazine Team4 min read
Key takeaways
- Data science is a team. Data engineers, data scientists and business stakeholders each play a part in turning data into decisions.
- Architects and engineers differ. An architect builds the platform, such as a Snowflake instance; the engineer designs how data lands in it.
- No statistics degree required. Some math is needed, but Boykin said you do not have to be a statistician.
- Start where you are. A path can begin as a business or data analyst and progress as skills grow.
- Learn Excel. Leaders often want answers in Excel, whatever tool the analysis used.
In a 2024 AI Connect session, Dr. Frances Boykin, Director of Advanced Analytics at AT&T, walked students and early-career professionals through the roles that make up a data science team and how to build a career in the field. Boykin, who said she has been at AT&T for about 18 years with a background in data engineering and data science, presented the talk as part of the AI Connect K-12 Data Science Conference 2024.
Her central point was that data science is a team effort. "The data scientist on their own" cannot deliver business decisions, she said, without the other roles that gather, move and apply the data.
What are the main roles in data science?
Boykin described three core groups: data engineers, data scientists and business stakeholders. Engineers provide the foundational infrastructure that lets an organization capture and use data. Stakeholders set the need, saying what information matters to the business. Data scientists take the raw data and stakeholder requirements and turn them into analyses, programming and "the stories that come from our data."
She tied the cycle to three ideas: data operationalization, data monetization and data optimization. Together, she said, they keep data flowing from collection through to business value.
How do business analysts, data analysts, architects and engineers differ?
Boykin went role by role. Business analysts improve business processes and act as intermediaries between the business and IT. Their favorite tools, she said, are Excel and visualization tools such as Tableau, and they are "conscious storytellers."
Data analysts collect and process data and do a small amount of statistical analysis using tools such as R, Python and SQL. Data architects build the infrastructure and the "blueprint" for data management, with skills in data warehousing and ETL, the process of pulling data from sources and landing it in one place.
Data engineers, a team she manages at AT&T, develop data sets for use across the organization and need database, modeling, ETL and API skills. Answering a question from the audience, she drew the line this way: an architect builds out a platform such as a Snowflake instance, while the engineer designs how data lands in it.
Do you need to be a statistician to be a data scientist?
No, according to Boykin. Many people assume data science demands a deep quantitative background, but she said AT&T does not have many true statisticians. Some math is required, she said: "you don't have to be a statistician."
Data scientists, in her description, hold a little of every other role: distributed computing, predictive modeling, visualization, math and machine learning. They need a curious mindset about what the data means. Database administrators keep databases running, and analytics managers, her own role now, connect the team's daily work to what company leaders will need for large decisions.
What is a realistic path into data science?
Boykin said you do not have to start as a data scientist. She began as a business analyst, moved into data analysis, then data engineering, architecture and data science, and now manages teams. Each step along that continuum, she said, requires more skills.
She showed a popular online roadmap covering fundamentals, machine learning, big data, statistics, visualization, data ingestion, data cleansing and text mining. It is useful for seeing what each area contains, she said, but not a checklist. Her advice was to "pick your poison": choose a focus, pick the topics that feel comfortable and lean into them.
Citing a DataCamp chart, she noted that 31% of data scientists held master's or professional degrees and 37% were college graduates, with others coming through technical schools or learning on the job. She suggested certificate programs have likely pushed the non-degree figures higher since.
What tools and experience should beginners get?
Boykin recommended getting hands-on experience: Kaggle, meetups, internships, boot camps, entry-level jobs and personal projects. Her advice was to "get in the data science scene" and follow the community, naming KDnuggets and Data Science 101 as favorites.
On tools, she listed R, Python, SAS, Tableau and Excel. AT&T uses Python, she said, and was once a SAS shop but no longer is. Excel remains essential because leaders rarely use the analyst's tool: "they're usually going to want their answers in Excel."
She also pointed learners to Python libraries for machine learning and visualization, including pandas, scikit-learn, NumPy, Seaborn and Matplotlib, and noted data science applies in fields from travel and marketing to healthcare and insurance.
Where can students learn more about data science?
Boykin said she planned to share her slides and resources with the organizers and invited questions through LinkedIn.
AI Connect continues to run free live webinars. The 2027 schedule includes 24 free live webinars across four series: the AI x Enterprise Panel Series, AI Demo Days, Agentic AI Showcases and the Plug & Play Series. Details are at /webinars/, and registration for 2027 events is coming soon.
Frequently asked questions
What is the difference between a data architect and a data engineer?
Boykin said the architect builds the platform, such as a Snowflake instance, while the data engineer designs how data lands in it and moves it between systems.
Do you need to be a statistician to be a data scientist?
No. Boykin said some math background is needed but you do not have to be a statistician, and AT&T has few true statisticians.
How can beginners get started in data science?
She suggested starting as a business or data analyst, picking focus areas from learning roadmaps, and building experience through Kaggle, meetups, projects, internships and boot camps.
This article is based on remarks by Dr. Frances Boykin, Director of Advanced Analytics, AT&T at AI Connect K-12 Data Science Conference 2024 on July 25, 2024. More coverage: More AI Connect Events.
Filed under: People, Culture & Leadership, AI Use Cases & Applications
