Why data analysis
Every company generates data and almost none of them can read it. That gap is the job, and it is one of the few skills that is genuinely learnable in months rather than years.
Analysts across Africa are working remotely for companies in the US and the UK. The tools are the same everywhere, and so is the interview.
Banking, fintech, e-commerce, health, telecoms, government. All of them generate data, all of them struggle to use it.
An analyst who knows the fundamentals and the AI shortcuts does the work of several who know neither. You learn both here.
Unlike engineering, which takes years, Excel and SQL are learnable and Power BI is visual. The market rewards speed to skill.
How the week runs
Not pre-recorded videos you are left alone with. A classroom that opens every day, on a timetable, with people who notice if you stop turning up.
Kemi teaches the day's lesson with slides, examples and a quiz at the end.
The next step, built directly on yesterday. Practice labs you do as you go.
The week's hardest idea, taken slowly, with worked examples.
You do it yourself. Guided exercises against real data, marked when you hand them in.
A working analyst, live, taking questions and demonstrating real workflows.
The whole week again, plus open questions and feedback on what you handed in.
Kemi is there at every other hour of the week as well: she explains, she quizzes you, and she marks your work the moment you submit it.
What you will be able to do
- Clean a dataset that arrives with duplicates, blanks and mixed date formats
- Write the Excel that does the work: SUMIFS, XLOOKUP, INDEX and MATCH, PivotTables
- Build a one-page Excel dashboard with KPI cells, charts and slicers
- Query a real MySQL database: joins across tables, subqueries and CTEs
- Say what a query result means in plain English a director can act on
- Model data properly in Power BI: a star schema with relationships that behave
- Write DAX that answers a question, including CALCULATE and filter context
- Ship four projects on real business data, the last one with no template at all
The syllabus
Everything you learn across the twelve weeks, split into three tool tracks. You work through each one with hands-on exercises and real datasets, and each track ends with a project. By the capstone you are using all three tools together on one business problem.
Excel 22 topics · weeks 1 to 4
Your way around Excel: the ribbon, worksheets, cells, ranges, and how Excel is structured as an analytical tool.
Text, numbers, dates and logical values. Knowing your data types is the first step before writing any formula.
Bringing external data in correctly through the Data tab. Most real data arrives as a CSV, so this is where every project starts.
Spotting problems before they ruin the analysis: duplicates, blanks, wrong types, inconsistent text, wrong formats.
Fixing messy text: extra spaces, capitalisation, and characters that do not belong.
Extracting and reshaping parts of a text value, for when data sits in one column but needs to be split.
Finding and removing duplicate rows. One of the most common cleaning tasks on any real dataset.
Fill, flag or exclude. The right answer depends on the column and on the business context.
Highlighting cells by rule so patterns and problems are visible without writing a formula.
Cell references, order of operations, absolute against relative, and how to read a formula you did not write.
The five foundational aggregations. You will use these in almost every analysis you ever do.
Making Excel decide: logic that categorises, flags or calculates differently depending on what is in a cell.
Several conditions, handled cleanly. IFS replaces nested IF when there are several possible outcomes.
Summing on one or many conditions. The most used analytical functions in Excel for business reporting.
Counting on conditions: how many transactions were above a threshold, how many customers are in a region.
Conditional averages, for comparing performance across categories, regions or periods.
Pulling values from one table into another on a key. The modern replacement for VLOOKUP and the one you will keep using.
The more flexible alternative, and when it is worth reaching for on large or awkward datasets.
Summarising thousands of rows instantly. The fastest way to answer questions about grouped data.
Turning a summary into charts and letting people filter it. This is where a static table becomes a dashboard.
Cleaning and reshaping before data reaches the sheet, and the difference between Power Query, formulas and DAX.
One clean, readable page that answers specific business questions with charts, KPI cells and slicers.
SQL 19 topics · weeks 5 to 7
The structure of a relational database, before a single query. Everything after this is built on it.
How tables link. Primary keys identify a record; foreign keys make the relationships that make SQL powerful.
Setting up your environment and confirming a real database is live before any analysis starts.
Your first query: what columns to return, and which table to look in. Every query starts here.
Filtering to the rows that meet a condition: equals, greater than, less than, not equal.
Sorting results, and returning only the unique values in a column.
Restricting how much comes back, and renaming columns so the output can be read and presented.
New columns written inside the query. Revenue is quantity times price; it does not need to be stored.
Missing data, dealt with properly, so a calculation does not silently break on it.
Conditional logic inside a query: categorise, label or calculate differently per row.
The five aggregations behind almost every business report you will write in SQL.
What turns individual transaction rows into a summary by region, category or month.
WHERE filters rows before aggregation, HAVING filters the groups after. Knowing which is which matters.
Cleaning and reshaping text without leaving the database.
Grouping by month, measuring the days between two events, filtering a period.
Combining rows from two tables where a match exists in both. The join you will use most.
Keeping everything on the left, and joining three or more tables to answer questions that span the database.
One query inside another, so a complex question can still be answered in a single statement.
The WITH clause: multi-step analytical logic that is easy to follow and easy to debug.
Power BI 16 topics · weeks 8 to 10
The report canvas, the data and fields panes, and the three views: Report, Data and Model.
Bringing in data from the sources you will actually meet at work.
What it is and why it exists: this is where data is cleaned, before the model, not after.
Duplicates, data types, nulls and inconsistent values, handled before anything loads.
Splitting, merging, replacing and creating custom columns so the shape fits the analysis.
Appending stacks tables on each other; merging joins them the way a SQL JOIN would.
What actually happens at Close and Apply, and where DAX and relationships take over.
The structure that connects your tables and makes your calculations correct rather than plausible.
The most important pattern in Power BI modelling. Facts hold transactions, dimensions hold the context.
Cardinality and cross-filter direction. Get these wrong and every measure you write answers wrongly.
The distinction that changes everything: stored row by row, or evaluated at query time under a filter.
Four functions that cover most of the basic calculations any dashboard needs.
The most important function in DAX. Changing filter context is the line between basic and analytical Power BI.
Choosing the right visual for the question, and configuring each one properly.
Letting people filter the view, and click from a summary into the detail underneath it.
Visual hierarchy and arrangement, so a business user finds the answer immediately.
The projects you will build
Four briefs, one at the end of each track and a capstone at the end of the programme. Every dataset is purpose-built, with realistic structure and deliberate data quality problems in it, because that is the kind of data a real analyst is actually handed. No real company data is used.
NovaBridge Retail: sales performance analysis
NovaBridge Retail is a Nigerian multi-branch chain selling electronics, fashion, groceries and household goods across five cities. You are the analyst. The head of commercial operations wants a sales performance review of the last twelve months: clean the data, analyse it, build the dashboard, and tell them what it means.
| Table | Records | Key columns |
|---|---|---|
| Sales transactions | 2,500 | Transaction_ID, Date, Customer_ID, Product_ID, Quantity, Unit_Price, Discount, City, Sales_Rep |
| Products reference | 80 | Product_ID, Product_Name, Category, Cost_Price |
| City reference | 5 | City_ID, City, Region, Branch_Manager |
Karya Logistics: delivery operations analysis
Karya Logistics runs last-mile delivery across eight cities. The operations director wants a review of delivery performance, rider efficiency, customer experience and demand patterns. You have the full database. Core querying, joins and aggregation, analytical SQL, and then the part most people skip: for every analytical answer, sixty words of plain English telling the director what it means and what to do about it.
| Table | Records | Key columns |
|---|---|---|
| orders | 8,000 | order_id, customer_id, rider_id, zone_id, order_time, delivery_time, status, delivery_minutes, rating |
| customers | 1,200 | customer_id, name, city, registration_date, segment |
| riders | 60 | rider_id, name, zone_id, hire_date, vehicle_type |
| zones | 8 | zone_id, zone_name, city, region |
| payments | 8,000 | payment_id, order_id, amount, payment_method, payment_date, payment_status |
Aura Telecom: customer and revenue intelligence
Aura Telecom is a mobile network operator with 50,000 subscribers across six states. The chief executive wants a dashboard the executive team opens every month: revenue, customer health, plan performance and service quality. You build the model and the dashboard from raw data. There is no template. You design it: an executive summary, revenue analysis, customer health and complaints, each page answering a named question.
| Table | Records | Key columns |
|---|---|---|
| Customers | 50,000 | Customer_ID, Name, State, Segment, Plan_ID, Registration_Date, Status |
| Plans | 12 | Plan_ID, Plan_Name, Monthly_Fee, Data_GB, Call_Minutes, Category |
| Payments | 180,000 | Payment_ID, Customer_ID, Payment_Date, Amount, Method, Status |
| Complaints | 8,500 | Complaint_ID, Customer_ID, Date, Type, Resolution_Status, Days_To_Resolve |
Orbis Health: patient and operations analytics
Orbis Health is a private hospital group with six branches across four states. The executive team wants a review of patient volumes, revenue, operational efficiency and doctor performance. You receive the raw data and the brief. Nothing else: no template, no walkthrough, no hints at the answers. You decide how to clean it, how to query it, how to model it and how to present it. It is the closest thing to a real analyst engagement you will do here, and it uses all three tools at once.
| File | Records | Columns |
|---|---|---|
| patients | 8,000 | patient_id, name, age, state, registration_date, insurance_type |
| appointments | 25,000 | appointment_id, patient_id, branch_id, doctor_id, appointment_date, department, status, wait_minutes |
| billing | 22,000 | billing_id, appointment_id, amount_charged, amount_paid, payment_method, insurance_claim, payment_date |
| doctors | 120 | doctor_id, name, department, branch_id, hire_date, specialisation |
| branches | 6 | branch_id, branch_name, state, capacity, opening_date |
What is wrong with the data, on purpose
Every dataset ships with real problems in it. Finding them is part of the work, because data that arrives clean is data somebody else already analysed.
| Dataset | Built in |
|---|---|
| NovaBridge Retail | Duplicate rows, inconsistent category names, blank discounts, negative quantities, mixed date formats |
| Karya Logistics | Null customer ratings, inconsistent zone capitalisation, duplicate delivery records, missing rider IDs |
| Aura Telecom | Null states, duplicate payments, inconsistent Plan_ID formats, missing registration dates |
| Orbis Health | Null insurance types, duplicate billing rows, inconsistent department names, missing doctor IDs, invalid wait times |
The tools you will use
Every one of these has a job in the analyst workflow. Nothing has been added to make the list look longer, and every one has a free version that covers the whole programme. You do not need to buy anything.
Cleaning, formulas, PivotTables, Power Query and Excel dashboards.
SQL querying and relational analysis across the whole SQL track.
Your SQL development environment: where you write and run every query.
A second SQL environment for selected exercises. Same concepts, different syntax in places, which is what real jobs look like.
Data modelling, DAX and the business intelligence dashboards.
A browser-based fallback for anyone who cannot install MySQL locally.
Your pod, and how the work is marked
You are placed with nine other people on this run, with a chat of your own, a pod league table and the same deadlines. The seat is earned rather than handed over: you are screened first and you say out loud what you are willing to give. It is the difference between studying alone and having nine people notice when you go quiet.
Every project is read by your tutor, marked out of 100 and returned with what to fix. The SQL brief asks you to write what your results mean in plain English, and that is marked too, because an analyst who cannot explain a number has not finished the job.
One brief, one submission, ten people. The pod builds it together and the grade pays every active member, which means the work has to be divided and somebody has to keep it moving. It is the closest thing here to a real analytics team.
Every class you attend and every piece of work you hand in pays Skill Points. They move you up the levels, they decide where your pod sits in the league, and they are part of what your certificate is issued against.
Who teaches it
Questions
No. This starts at zero, with what a spreadsheet is, and every topic builds on the one before.
A laptop that runs Excel, MySQL Workbench and Power BI Desktop, and an internet connection good enough for a live class.
No. Every tool on the programme has a free version that covers everything you need, and there is a browser-based SQL fallback if you cannot install MySQL locally.
About eight hours a week: one to two hours a day, plus the live classes. It is built to be done alongside a job.
Four projects: a sales performance review in Excel, a delivery operations analysis in SQL, a telecoms dashboard in Power BI, and a healthcare capstone using all three.
The structure is. Every dataset is purpose-built, with deliberate quality problems in it, so you learn on the kind of data an analyst is actually handed. No real company data is used.
The Saturday class reviews the whole week, and Kemi is there at any hour to go back over anything you missed.
Yes, a verifiable Webbo3 certificate, earned by finishing the curriculum and the projects rather than by paying.
It gives you the skills, four real projects and the job-search work at the end. The applying is still yours to do.
Yes. The classes run online and the community is across Africa.
Apply for this course
Applying costs nothing and takes about two minutes. Three days of the platform are free before you pay anybody anything.
Apply for this courseApplying is free. You choose how to pay on the next screen, and your tutor decides on your place.