Building a Data Mindset: How to Think Like a Data Scientist

WebinarSeptember 2024

Data analysts, data scientists, and other data professionals are often key employees of an organization, and it's likely that at some point you'll need to collaborate with one of them. If you want to collaborate effectively with a data professional, you'll need to learn to think like them.

In this session, Dave Wentzel, a Technical Evangelist at Microsoft, helps you understand the data mindset so you can have an intelligent conversation with a data scientist, and perhaps try analyzing some data yourself. You'll learn core data jargon, see the data science workflow in action, and see how data can be used to solve business and real-life problems.

Key Takeaways:

Learn to talk like a data scientist by understanding data jargon.
Understand simple data science workflows.
Improve your data literacy for better collaboration with data professionals.

Resources

Slides

Dave's GitHub Repo

Summary

In the exciting field of data science, essential technical abilities like Python and machine learning are key, but equally important is the mindset that steers effective data use for decision-making. This session, led by Dave Wentzel, a technical evangelist at Microsoft, explored how data can solve business and science problems, particularly highlighting the significance of intuitive understanding beyond mere technical knowledge. Key discussions included the analytics maturity model, which outlines the progression from descriptive to prescriptive analytics, and the significance of design thinking in resolving complex business problems. Dave also shared insights on the pitfalls of traditional data modeling and emphasized the value of outlier analysis in uncovering important insights. The session concluded with practical advice on developing data literacy and the importance of continuous learning and feedback in the data science process.

Key Takeaways:

Data science requires a mindset beyond technical skills, emphasizing intuitive understanding and decision-making.
The analytics maturity model progresses from descriptive to prescriptive analytics, requiring increasing levels of data sophistication.
Design thinking is important for addressing complex business problems, allowing for creative and collaborative solutions.
Outlier analysis can reveal critical insights that traditional data approaches might miss.
Continuous learning and feedback are essential for developing data literacy and enhancing data-driven decision-making.

Deep Dives

The Data Scientist Mindset

Understanding the mindset ...
Read More

of a data scientist goes beyond mastering technical skills such as coding in Python or building machine learning models. It involves adopting an all-encompassing strategy to problem-solving that emphasizes intuition and decision-making. Dave Wentzel, with 20 years of experience, highlighted the importance of collaboration between data scientists and business professionals, noting that effective communication is key to identifying and addressing the correct business problems. He shared a personal anecdote about early career mistakes, emphasizing how important it is to align technical solutions with business needs. "Communication skills are incredibly powerful," he stated, emphasizing the need for data scientists to not only understand data but also to convey insights effectively to stakeholders.

Analytics Maturity Model

The analytics maturity model is a framework that divides data analysis into four levels: descriptive, diagnostic, predictive, and prescriptive analytics. As explained by Dave, moving up this ladder requires increasingly sophisticated data handling capabilities. Descriptive analytics, which provides historical insights, is the foundation. Diagnostic analytics goes deeper to understand the 'why' behind the data. Predictive analytics uses historical data to forecast future outcomes, a task often associated with data scientists. However, Dave pointed out that the real value lies in prescriptive analytics, which not only predicts outcomes but also suggests actions to optimize results. This progression is not solely technical but requires a strategic mindset to leverage data effectively for business outcomes.

Design Thinking in Data Science

Design thinking, a methodology for creative problem solving, is integral to data science, particularly in addressing complex business problems. Dave Wentzel emphasized its importance in bringing diverse teams together to collaboratively tackle 'fuzzy' or 'squishy' problems—those without clear-cut solutions. By involving stakeholders from IT, business, and data science, design thinking encourages diverse perspectives and innovative solutions. Dave shared examples of rapid prototyping sessions where design thinking was employed to redefine business questions and explore data collaboratively. "It's about getting people in a room and debating things as we look at data together," said Dave, highlighting the value of this approach in creating meaningful and actionable insights.

The Importance of Outlier Analysis

Outlier analysis is a vital component in data science, often revealing insights that conventional methods might overlook. Dave Wentzel shared an interesting case study involving a convenience store's failed ice cream sales strategy, which was initially based on temperature data. By examining outliers, they discovered that the sales pattern was actually influenced by school schedules, not solely weather conditions. "The interesting data exists in the outliers," Dave explained, illustrating how outliers can uncover hidden variables or confounders that significantly impact results. This example emphasizes the importance of questioning assumptions and thoroughly examining all aspects of the data to obtain a comprehensive understanding.