Career change

Data analyst to data scientist resume

Data analysisData science

Data analysts already clean data, define metrics, and explain results to people who make decisions. Data science adds more modeling, experimentation, and programming depth. This page shows how to describe your work in that language, which gaps hiring managers will test, and how to build evidence without exaggerating what you have shipped.

What carries over

The skills you have, and the ones to add

YouSkills you likely already have

  • SQL
  • BigQuery
  • Python
  • Excel
  • Looker
  • Tableau
  • Power BI
  • Data modeling
  • Experiment analysis
  • Cohort analysis
  • Forecasting
  • Data quality

From the Data analysis guides. A tick marks skills the new role also lists.

NextSkills to build and show

  • Machine learning
  • Experiment design
  • Causal inference
  • Model evaluation
  • NLP
  • Ranking
  • PyTorch
  • scikit-learn
  • Spark
  • Feature engineering

From the Data science guides. List only what you can back with real work.

SQL fluency, metric definition, data cleaning, dashboarding, and stakeholder communication all carry over. So does your sense of which questions are worth answering. Keep tools like Looker or Tableau, but shorten that part and expand anything involving Python, testing, or prediction, since that is where the new role will focus.

Making the case

Why the move makes sense

Why it works

A strong analyst understands the business question, the quality of the data, and what a number can and cannot say. Those are the hardest parts of data science to teach. If you have used Python, tested hypotheses, or built forecasts, you are already close. Many data scientists began as analysts, because domain knowledge and trust with stakeholders carry real weight.

What readers will question

Hiring managers will question statistics and machine learning depth. They will ask whether you can choose and evaluate a model, avoid leakage, design a valid experiment, and write code others can maintain. Dashboards and SQL queries do not prove those skills by themselves. They will also check whether you can explain a model's limits to non-technical colleagues.

How to bridge it

Bridge it with a small number of real projects that include modeling or experiment design, even if modest. A churn estimate, a demand forecast, or an A/B test analysis with a clear method works. Write down the question, the approach, the evaluation, and the decision it affected. Add formal learning in statistics if gaps are real, and say what you built, not what you studied.

Reframe your bullets

Same work, new language

Each pair describes the same underlying work. The second leads with the result, the measure and the action, the way a Data science reader expects.

Your current resume

Built dashboards tracking weekly signups and activation for the product team.

Reframed for Data science

Modeled signup data in Python to identify onboarding drop-off, then recommended a simplified first step that lifted activation 9% in a controlled test after launch.

A reporting task becomes a tested recommendation, which is closer to the scientist's work.

Your plan

Three steps to a convincing resume

  1. 1Pick one modeling project

    Choose a real question from your workplace data, or a public dataset, and build a simple model with a held-out test. Write the question, method, result, and limits in one page. Keep it honest and small enough to finish.

  2. 2Strengthen statistics and Python

    Review hypothesis testing, regression, and model evaluation, then practice in Python with scikit-learn or a similar library. Use your own SQL knowledge to prepare data well. Note courses by name only if they led to something you built.

  3. 3Reframe your current work

    Ask to own an experiment, forecast, or segmentation at work, and document it carefully. Then rewrite your bullets around questions, methods, and decisions instead of dashboards, and compare the structure with the data scientist example.

Common slips in this move

  • Listing every library you touched without showing a project where you used them.
  • Calling dashboard work machine learning to match the job description.
  • Reporting model accuracy with no mention of validation or the baseline.
  • Hiding your analyst strengths instead of using them as your clear advantage.
Guides to read

Study both sides of the move

Writing every bullet this way? Read the XYZ formula.

Questions

FAQ

Do I need a master's degree to become a data scientist?

Some employers prefer one, especially for research-heavy roles, but many hire analysts who have shown strong applied work. Read the job descriptions you want and note what they actually ask. Evidence of modeling, experiments, and clear communication often counts more than a credential, though requirements vary by company and team.

How much machine learning do I need to show?

Enough to prove you can frame a problem, build a reasonable model, evaluate it properly, and explain the limits. One or two well-documented projects are better than a long list of techniques. Product data science roles often emphasize experiments and causal thinking over complex models, so match your evidence to the role.

Ready to write it?

Start from a resume built for the role you want.

Open the Data science guide for a complete one-page example and download it as a PDF or DOCX.

Open the Data science guide