📊 Data Science Methodology 101: From Problem to Approach – Business Understanding
Understanding Business Needs: A Step-by-Step Data Science Approach
Imagine this scenario:
You’re called into a meeting with your boss. There’s an urgent task with a tight deadline, and you both go over the details to ensure everything is on track. The meeting ends on a confident note.
But later that day, as you dive deeper into the problem, you realize you have more questions than answers. The boss isn’t available until tomorrow—should you move forward blindly or wait for clarification?
👉 In data science, the right choice is always to clarify first.
This first step is called Business Understanding, and it sets the foundation for the entire data science methodology.
🔹 Why Business Understanding Matters
Before analyzing data, writing code, or building models, a clear understanding of the problem is essential. John Rollins, IBM Senior Data Scientist, emphasizes that a clearly defined question guides the analytic approach.
Too often, data teams answer the wrong question. The methods might be perfect, but if the question itself was misunderstood, the solution won’t solve the real business problem.
🔹 From Goals to Objectives
Business understanding starts with identifying the goal of the person asking the question.
For example, if a business owner asks:
👉 “How can we reduce the costs of performing an activity?”
The real question is:
Is the goal to improve efficiency?
Or to increase profitability?
Once the goal is clarified, you can break it into objectives. Structured discussions with stakeholders help prioritize these objectives and build a realistic plan of action.
🔹 Case Study: Healthcare Budget Allocation
Let’s look at a real-world example.
Problem:
An American healthcare insurance provider faced cuts in public funding for patient readmissions. To avoid raising insurance rates (an unpopular option), they needed to find a way to maximize their limited healthcare budget.
Step 1 – Clarify the goal:
The team worked with IBM data scientists to define the main objective. They discovered that patient readmissions were a key area of concern.
Step 2 – Break down the problem:
30% of rehab patients were readmitted within one year.
50% were readmitted within five years.
Patients with congestive heart failure were the top contributors to readmissions.
Step 3 – Translate into business requirements:
The project sponsors identified four must-haves for the analytics model:
Predict readmission outcomes for congestive heart failure patients.
Predict readmission risk levels.
Identify combinations of factors leading to readmissions.
Provide an easy-to-understand process for predicting new patient risks.
Step 4 – Involve stakeholders:
The insurance company involved healthcare authorities and IBM scientists. They conducted an on-site workshop with business sponsors to set direction, stay engaged, and provide ongoing support.
🔹 Lessons Learned
The Business Understanding stage ensures:
The right problem is defined.
Goals and objectives are clarified.
Stakeholders are engaged early.
Requirements for the final model are clear.
Skipping this step often leads to wasted effort and irrelevant insights.
🔹 Wrapping Up
Business understanding is the first stage of the Data Science Methodology. It answers the question:
👉 “What is the problem we are trying to solve?”
With this foundation, data scientists can confidently move to the next stage: the Analytic Approach.
Stay tuned for the next article in this series, where we’ll explore how to select the right analytic approach once the business problem is crystal clear.
Source: IBM Data Science Professional Certificate (Coursera)
Instructors: A special thanks to the incredible IBM instructors and the entire IBM Skills Network Team.
Disclaimer: This blog post is part of my personal learning journal where I document my progress through the IBM Data Science Professional Certificate. These articles represent my personal understanding and interpretation of the course material. They are not official course notes and are not endorsed by IBM or Coursera.