Data Analyst vs Data Scientist vs Data Engineer in Kenya

McTaba runs coding and data bootcamps in Kenya. This guide is written to be useful whether or not you ever join one.

Picture a Nairobi lender that holds ten years of loan repayments. Three different people work on that data every morning. One writes the Monday report on repayments. One builds a model that scores new borrowers. One makes sure the repayment records arrive in the warehouse at all.

Those are the analyst, the scientist and the engineer. Job boards blur the titles, but the three jobs are different. This guide reads them from real Kenyan job ads, so you can choose before you spend months learning the wrong toolkit. If you want the wider career map first, read the data science career guide for Kenya.

The ads we quote were on MyJobMag Kenya on 3 November 2026. We read all of them in full, and the method is in our count of what Kenyan employers ask for.

Data analyst: what happened, and why

The analyst turns records into answers for managers. A Kenya Airways analyst ad asks for someone to analyse business data for heads of department, spot trends in large data sets, and audit system data for accuracy and duplicates (Kenya Airways data analyst ad, checked on 3 November 2026). At a lender, a junior ad from Pezesha lists SQL queries, repayment-rate reports and customer segmentation (Pezesha ad, checked on 3 November 2026).

Daily tools: Excel, SQL, Power BI or Tableau, sometimes Python. Output: a dashboard, a weekly report, an answer in a meeting.

Data scientist: what happens next

The scientist builds models. A Gulf African Bank ad lists predictive models for credit risk, fraud and anomaly detection, customer churn and collections (Gulf African Bank ad, checked on 3 November 2026). A Co-operative Bank ad asks for models with benchmarks, validation, monitoring, fairness and explainability, deployed with engineers (Co-operative Bank ad, checked on 3 November 2026).

Daily tools: Python, SQL, scikit-learn, sometimes R. Output: a model that sits behind a loan app or a fraud alert, and a note on how well it works.

Data engineer: the pipes

The engineer moves and cleans the data so everyone else can trust it. Watu Credit's ad lists ingestion pipelines into a data warehouse, dbt and SQL clean-up of raw tables, and data quality and governance (Watu Credit ad, checked on 3 November 2026). Victory Farms asks for pipelines from its ERP into a Snowflake warehouse, with automated checks for missing data and mismatches (Victory Farms ad, checked on 3 November 2026).

Daily tools: SQL, Python, cloud warehouses, pipeline tools. Output: tables that refresh on time and do not break.

	Data analyst	Data scientist	Data engineer
Main question	What happened?	What will happen?	Is the data right and on time?
Core tools in the ads	SQL, Power BI, Excel, Tableau	Python, SQL, R	SQL, Python, cloud tools
Middle years asked (lower bound)	2	3.5	4
Ads in our count (of 62)	26	15	19
Paylab monthly gross, 80% range (KES)	40,231 to 155,656	51,664 to 195,546	no page found
Payscale average base, per year (KES)	504,822 (130 profiles)	380,000 (60 profiles)	1,210,095 (11 profiles)

Sources: MyJobMag searches for the ads and years, Paylab data analyst, Paylab data scientist, Payscale data analyst, Payscale data scientist and Payscale data engineer, all checked on 3 November 2026.

Read the pay rows with care. Paylab's data scientist page says it has fewer than 20 verified respondents, and its analyst page has 25. Payscale's engineer average rests on 11 profiles. The two sites disagree sharply, which tells you the market is thinly measured, not that one is right. We unpack this in the data analyst salary page and the data scientist salary page.

If you want one programme that touches all three roles in its first months, see the Data Science to AI Engineering syllabus.

Analyst: the widest door. It had the most ads (26 of 62) and the lowest years asked. Six ads were internships or junior posts, and four of the 62 ads accepted a diploma. Build one dashboard on real data and learn SQL first. Our data analyst career guide goes deeper on the route.

Scientist: the maths door. Ads ask for statistics and model validation, and the middle ad wants three to four years. A common route is an analyst job first, then a move into modelling. A degree in statistics, maths or computer science helps. Our statistics guide for data science covers the base.

Engineer: the software door. Ads ask for pipelines, cloud warehouses and code that does not break at 3 a.m. Many engineers come from software development. If you already code, this is the shortest hop.

We found no Kenyan telco data ad on the day we looked, so we do not claim what telco teams want. The bank, insurer, lender and airline ads above are the evidence we can point to.

These three phrases get used as if they meant one thing. They do not. Imagine one Safaricom M-Pesa statement, a PDF full of transactions, and three questions asked of it. These are examples, not market data.

• Data analysis: "How much did this customer spend on airtime each month, and where did the spending jump?" You clean the rows, group them, count them, and report. Output: a table and a sentence.
• Data visualisation: "Show the owner those monthly totals so she sees the jump in two seconds." You turn the table into a line chart or a dashboard. Output: a picture that carries the answer. It is one tool of the analyst, not a separate career in most Kenyan ads.
• Data science: "Given six months of statements, will this customer repay a KES 20,000 loan?" You build and test a model on past cases, then check it on new ones. Output: a prediction with a known error rate.

Analysis explains the past. Visualisation makes the explanation easy to see. Science predicts, and must prove the prediction holds. Our guide to AI, machine learning and data science places all three in one map, and the data storytelling guide covers the visual side.

• You like answering a manager's question and seeing the chart land: analyst.
• You like maths puzzles and can sit with an error for a week: scientist.
• You like fixing things that break and already write code: engineer.
• You are changing careers from a non-technical job: start as an analyst, then decide.
• You are a software developer: consider engineering first, then modelling.

Not sure yet? Is data science worth it in Kenya? sets out the costs and the odds, and where data jobs are posted shows who is hiring.

Our Data Science to AI Engineering programme starts on 11 January 2027 and costs KES 140,000, over 5 months full time or 8 months part time. Its phases include Python, SQL, Power BI, machine learning and MLOps, so it covers the three roles at an entry level. The first cohort has not started, so we have no graduate outcomes to quote, and no school can promise a job. Read the Data Science to AI Engineering syllabus and judge it against the ads above.