A data analyst looks at data a business already has and explains what happened: which product sold most last month, where customers drop off, why costs went up. A data scientist uses the same data to build models that predict or decide something: which customer is likely to leave, what demand will look like next month, which transaction looks like fraud. The two jobs overlap a lot, and the titles get used loosely, so it helps to look at what the work involves rather than the name on the ad.
The two roles side by side
| Data analyst | Data scientist | |
|---|---|---|
| Main question | What happened, and why? | What will happen, and what should we do? |
| Daily tools | SQL, Excel, Power BI or Tableau, some Python | Python, SQL, machine learning libraries, sometimes cloud tools |
| Typical output | Dashboards, reports and a clear summary for a manager | Models, predictions and experiments, often built into a product |
| Maths needed | Basic statistics: averages, percentages, spread, simple tests | More statistics, plus how machine learning models learn and fail |
| Common first step | Often the first data job people get | Often reached after time as an analyst or in a technical degree |
What a data analyst does in a week
Most of an analyst's week goes into getting data out of a database with SQL, cleaning it so the numbers can be trusted, and turning it into something a manager can read in two minutes. That might be a Power BI dashboard that refreshes every morning, or a short note that says sales in one branch fell because a product was out of stock for ten days. Cleaning takes longer than people expect. Real data has missing values, spelling differences and duplicates, and a dashboard built on dirty data is worse than none.
The other half of the job is explaining. An analyst who finds the right number but cannot explain it clearly has only done half the work.
What a data scientist does in a week
A data scientist does the same cleaning and querying, then goes further. They choose a model, train it on past data, test it on data it has never seen, and check whether it is good enough to trust. They spend a lot of time on questions like: is this model accurate only because the test was too easy? Will it still work next month? What happens when it is wrong? In some companies a data scientist also works with engineers to put the model into an app.
Which of our courses fits each path
Our Data Science and Machine Learning course is built around the analyst path. Over 12 weeks it covers Python from zero, cleaning data in pandas, SQL including joins and window functions, statistics, a full week of Power BI and Streamlit dashboards, then starter machine learning. You finish with analysis and dashboards you can show, plus a first taste of models. If you are new to data, start here.
If you already know you want to build models and AI systems, look at AI Engineering and Machine Learning. It also starts with Python from zero, then spends more of its 12 weeks on machine learning, deep learning and large language models, and ends with a deployed AI system.
Neither course needs prior coding or maths. Both start from the basics.
Which one should you aim for first?
For most beginners, analysis is the more realistic first goal. The skills are the base of data science anyway: you cannot build a good model on data you do not understand. Once you are comfortable with SQL, cleaning and explaining numbers, moving towards modelling is a smaller step.
If you enjoy building things in code more than explaining numbers to people, the AI Engineering route may suit you better from the start.
The week-by-week syllabus for both courses is on each course page, and you can visit our classroom in Old Baneshwor to ask a trainer.
