Data analysts and data scientists both work with data, often use SQL and Python, and sometimes belong to the same team. Yet the two titles are far from interchangeable, even if companies use them very freely. A "data scientist" ad can describe a reporting job, and a "data analyst" ad an advanced modelling job. For a candidate, confusing the two leads to poorly targeted applications and disconcerting interviews. This article compares the tasks, tools, education and interviews, then shows how to read a job ad to recognise the real role and adapt your resume.
Two jobs, blurred titles
At the outset, the distinction is fairly clear. The data analyst answers questions about what happened and why: they explore data, build indicators and dashboards, and help teams make decisions. The data scientist is more interested in what will happen and what can be automated: they build forecasting, classification or recommendation models, and design experiments to measure effects.
Onisep’s profiles of the data analyst 🔗 and the data scientist 🔗 reflect this difference: the first stresses analysis and reporting to the business, the second modelling and algorithms.
In practice, three factors blur the boundary.
- Company size. In a small team, one person often does both: a dashboard on Monday, a forecasting model on Thursday.
- The appeal of the title. "Data scientist" is a prestigious title, used by some companies to attract candidates to mainly analytical roles.
- The emergence of intermediate jobs. Analytics engineer, product analyst, decision scientist: these titles cover varied mixes of the two jobs.
The title alone is therefore never enough. It is the content of the ad that tells you which role it is.
The tasks compared
| | Data analyst | Data scientist | |---|---|---| | Typical question | What happened, and why? | What will happen, what can be automated? | | Deliverables | Dashboards, ad hoc analyses, metric definitions | Models, experiments, prototypes, sometimes services in production | | Stakeholders | Business teams, management, product | Product, engineering, business teams | | Pace | Frequent requests, short deadlines | Longer projects, iterations | | Measure of success | Informed decisions, reliable and adopted metrics | Model performance, measured effect on the business |
A data analyst spends much of their time understanding teams’ questions, writing queries, checking data consistency and presenting results. A data scientist spends more time preparing datasets for training, testing models, evaluating them and working with engineers to put them into service.
Both jobs share one requirement: data quality. Without reliable data, neither the analysis nor the model has any value. That is why both often work closely with the data engineer, whose role is described in our article on the data engineer job.
Tools and skills
The tools overlap widely; it is how they are used that differs.
The data analyst primarily uses:
- SQL, intensively, to query the data warehouse;
- a visualisation tool: Power BI, Tableau, Looker, Metabase;
- a spreadsheet, always present in exchanges with the business;
- Python or R for more advanced analyses, increasingly expected;
- basic descriptive statistics and tests to interpret differences correctly.
The data scientist primarily uses:
- Python, with data manipulation and learning libraries (pandas, scikit-learn, XGBoost or LightGBM, sometimes PyTorch);
- SQL, to extract and prepare data;
- inferential statistics and experimentation: A/B tests, causal inference;
- tracking and production tools: MLflow, orchestrators, containers, depending on the level of industrialisation expected.
On the non-technical side, the data analyst is particularly expected to communicate well and understand the business; the data scientist to show methodological rigour and the ability to frame a poorly defined problem. Both must be able to explain a result to someone who does not know the technique.
Education and career paths
Both jobs are accessible through similar routes, with differences in level.
- Data analyst: business or engineering schools with a data specialisation, master’s degrees in statistics, econometrics or computer science, but also professional bachelor’s degrees, bootcamps and career changes from jobs that already involved figures (management control, marketing, logistics). The job remains open to self-taught profiles who master SQL and a visualisation tool.
- Data scientist: most often a five-year quantitative degree, engineering schools, statistics schools, master’s degrees in applied mathematics or machine learning, sometimes a PhD for the most research-oriented roles.
Many data scientists started as data analysts, and it is a natural path: you learn the company’s data and the business’s expectations before building models.
Reading a job ad to recognise the real role
To find out which role it really is, ignore the title and read the tasks and tools. A few clues:
Clues to a mainly analytical role, even under the title "data scientist":
- "build and maintain dashboards", "track performance indicators";
- Power BI, Tableau or Looker first in the tools;
- "answer ad hoc requests from teams";
- no mention of models, experimentation or production.
Clues to a modelling role, even under the title "data analyst":
- "develop forecasting models", "design A/B tests";
- scikit-learn, XGBoost, PyTorch, MLflow in the tools;
- "work with engineers to industrialise";
- prerequisites in inferential statistics or machine learning.
Clues to a hybrid role: a mix of both, often in a small team. It is an excellent learning ground, provided you know what you want to develop.
When in doubt, ask during the screening call: "What share of the time goes to dashboards, ad hoc analyses and modelling?" The answer will tell you almost everything. Interviews for modelling roles also include specific tests, described in our article on data scientist interview questions.
Adapting your resume to one or the other
The same background can be presented differently depending on the target role, without inventing anything.
For a data analyst role, highlight:
- the metrics you defined and the dashboards you built, with their users;
- the analyses that led to a decision;
- your mastery of SQL and of the visualisation tool in the ad;
- your ability to work with business teams.
For a data scientist role, highlight:
- the models you built, with the business question, the data, the validation method and the measured effect;
- the experiments you designed and analysed;
- your practice of Python and learning libraries;
- what was deployed or used beyond your own workstation.
The data scientist resume template illustrates the second presentation, with projects described by their problem, data and effect. The data engineer resume template shows, by contrast, a profile centred on data flows.
The sector matters too. A data analyst in banking or capital markets works with particular data and regulatory constraints: the specifics of a data analyst resume in finance 🔗 are covered in a separate guide, which details the tools and vocabulary expected in that sector.
Once you have chosen your positioning, the summary and the order of your bullets do most of the work. The TechCV editor lets you keep two versions of your resume, one focused on analysis, the other on modelling, and produces a PDF that recruitment software reads without errors. You can create your resume and duplicate the base version for each target.
Frequently asked questions
Can you move from data analyst to data scientist?
Yes, it is a common progression. It involves deepening inferential statistics, machine learning and Python, and taking on first modelling projects, ideally in your current job. Your knowledge of the company’s data and business will be an asset.
Do you need to code to be a data analyst?
SQL is essential. Python or R are increasingly expected, without requiring a data scientist’s level. A data analyst who automates their analyses in Python stands out clearly.
Which job hires more beginners?
Data analyst roles are generally more accessible at the start of a career, as they require less specialisation in modelling. Junior data scientist roles exist, but often require a five-year quantitative degree or prior experience.
Does the job title matter for the rest of your career?
Less than the content of the work. A recruiter will read your experience to find out what you actually did. Describe your tasks precisely, and you will be able to apply for both types of role if your background justifies it.