Data Analyst to Data Scientist

Why this transition works

• ✓Data analysts already understand the business domain and stakeholder needs — half of data science work
• ✓SQL and data manipulation are table-stakes skills already in place
• ✓Statistical literacy from analyst work transfers directly to ML feature engineering
• ✓Internal transitions at the same company have high success rate

The hard parts (don't skip these)

• !ML modelling depth requires deliberate study — Python, scikit-learn, deep learning basics
• !Production deployment (MLOps) is often the gap — analysts don't typically deploy systems
• !Statistical rigour for hypothesis testing and inference goes beyond analyst-level statistics
• !Some senior DS roles still gate-keep PhDs; mid-level transitions are more accessible

Step-by-step plan

1. 1

   Build Python and ML fluency

   Python with scikit-learn and pandas is the foundation. Coursera ML specialisations (Andrew Ng) or fast.ai practical courses are credible signals. Aim for 6-9 months of deliberate study.

2. 2

   Build 2-3 portfolio ML projects

   Real-data projects showing the full cycle: problem definition, data preparation, modelling, evaluation, written conclusions. Kaggle competitions are useful but original projects are stronger.

3. 3

   Run an ML project at your current company

   Internal credibility multiplier. A churn prediction model, a forecasting model, or a recommender — anything that shows you can ship ML work, not just train notebooks.

4. 4

   Internal transition first

   Talk to data science leadership at your company. Many DS teams welcome analysts who've built ML projects internally. Internal success rate is meaningfully higher than external.

5. 5

   External transitions: target product DS roles

   For external moves, target product data scientist roles (vs research DS roles). The product DS bar is more accessible to analyst transitions; research DS often requires PhD.