How to learn Power BI from scratch, and when to use Power BI or Excel

Power BI or Excel: which one you need

They compete less than it seems. Excel is a flexible spreadsheet: good for exploring, running a one-off calculation, building a budget or reviewing a list. Power BI is a reporting tool: it connects sources, defines indicators once and publishes them so other people can check them up to date. If you do the analysis yourself and for yourself, Excel is usually enough. If several people ask for the same report every week, that is a job for Power BI.

In client projects, the most common pattern is this: a monthly report built in Excel from exports of several systems, which takes days and reaches each department in a different version. Moving it to Power BI does not remove Excel from the company, but it does remove that cycle. The comparison table below sums up what each one does best.

What you need to know before you start

You do not need to code. It does help to arrive with solid Excel: knowing what a table is, having used pivot tables and feeling comfortable with functions like SUMIFS or XLOOKUP. If you also know Power Query in Excel, you are well ahead, because Power BI uses the same engine to prepare data. What almost nobody brings is the idea of a data model, and that you learn along the way.

Power BI Desktop is free and runs on Windows. Publishing and sharing reports requires licences for the online service, but you do not need to pay anything to learn and practise on your own computer. If you use a Mac, you will need a virtual machine or a Windows computer, something worth sorting out before buying a course.

What to learn and in what order

The most frequent mistake is starting with the charts, because they are the most eye-catching part. The order that works is the reverse: data first, then the model, then the calculations and finally the design. A pretty report on a badly built model produces numbers nobody can explain. At about five hours a week, this is a rough roadmap you can shorten or stretch depending on your time:

• Weeks 1 and 2: the interface, loading data from Excel and CSV, and basic cleaning with Power Query.
• Weeks 3 and 4: the data model, one-to-many relationships, fact and dimension tables, and a calendar table.
• Weeks 5 to 7: DAX. Measures with SUM, CALCULATE and FILTER, and time intelligence functions.
• Weeks 8 and 9: report design, interactions between visuals, slicers and navigation.
• Week 10 onwards: publishing, scheduled refresh, row-level security and connecting to databases.

The data model is the part you should not skip

People coming from Excel tend to build one huge table with every column, because that is how pivot tables work. In Power BI you pay for that with measures that do not add up and slow reports. The recommended approach is the star schema: a fact table with the transactions, such as sales or orders, surrounded by dimension tables like customers, products, branches and a calendar.

Understanding this changes how you write DAX, because many formulas that looked complicated become simple once the model is right. If a course covers modelling in a single ten-minute video, look for another one or supplement it with material dedicated to the topic. In the dashboards we build for clients, most of the time goes into the model and data quality, not the charts.

Practice projects

Pick a project with a concrete question and someone who could use the answer, even an imaginary someone. A dashboard without a question ends up as a collection of charts. For data, use your work data if you are allowed to, public sales or government datasets, or a sample database with orders and customers. These ideas cover every stage of the roadmap:

• A sales dashboard with revenue by month, comparison against the previous year and best-selling products.
• Receivables tracking: overdue invoices by customer and by age bracket.
• An expense report by department built from several Excel files combined in Power Query.
• Indicators for an operational process, such as response times or orders shipped on time.
• Rebuild in Power BI a report that is currently put together by hand in Excel, and compare the effort.

Common mistakes

Almost every beginner mistake in Power BI comes from carrying over Excel habits. They are not tool failures but approach problems, and they are quick to fix once you recognise them. These are the ones we see most when reviewing reports built by internal teams, and it pays to keep them in mind from your first project:

• Using calculated columns for everything when a measure is the right tool.
• Working without a calendar table and then fighting with period comparisons.
• Putting twenty visuals on one page instead of answering three questions well.
• Cleaning data by hand in Excel before loading it, instead of doing it in Power Query.
• Connecting to files saved on your desktop, which blocks automatic refresh.

How to choose a Power BI course

Power BI gets updates every month and the interface changes often, so the course’s last update date matters more than with other tools. A course from several years ago still works for DAX and modelling, but it can confuse you with menus that no longer exist. Check that the syllabus gives full sections to Power Query, the data model and DAX, not just to visuals.

On language, there are good Spanish-language courses to get started; for advanced DAX there is more material in English. If you want a certification, Microsoft has an official Power BI data analyst exam, and some courses follow its syllabus. A certification does not replace a portfolio, but it does give structure to what you need to study.

• Recently updated, with the interface you will see when you open Power BI.
• Sections dedicated to the data model and DAX, not only to charts.
• Downloadable data files to follow every exercise.
• A complete end-to-end project, from loading data to publishing.

When Power BI is no longer enough

Power BI is a tool for reading data, not for capturing it. If the real problem is that information is recorded in scattered spreadsheets, with errors and no record of who changed what, a dashboard just puts those errors in nicer charts. At TintoDev we see this often: the client asks for a dashboard and what they need first is a system where data goes in correctly.

It also falls short when the report has to live inside your own application for external customers, when data grows to the point of needing a data warehouse, or when you need logic beyond showing indicators, such as sending alerts or triggering processes. In those cases Power BI can remain the visual layer, but backed by a well-designed database and automated processes.

Where the courses are

Frequently asked questions

Do I need to know Excel before learning Power BI?

It is not required, but it helps a lot. If you already handle pivot tables and Power Query in Excel, you start with much of the ground covered. If you are starting from zero, spend a few weeks on Excel first, or at least on understanding how data is organised in tables.

How long does it take to learn Power BI from scratch?

At about five hours a week and with an Excel background, you can publish a useful dashboard in one or two months, as a rough guide. Writing DAX fluently and designing complex models takes several more months of practice with real data.

Does Power BI replace Excel?

No. Power BI replaces the recurring reports that are built by hand in Excel, but Excel is still better for one-off analysis, quick calculations and planning. In most companies the two coexist, and many users export data from Power BI to Excel to keep working on it.

Is Power BI free?

Power BI Desktop, the application for building reports, is free and is enough to learn. Publishing and sharing reports with other people requires licences for Microsoft’s service, and it is worth checking the terms on the official site because they change over time.

What is DAX and is it hard to learn?

DAX is Power BI’s formula language for creating measures such as totals, percentages or period comparisons. Basic formulas look like Excel ones; the hard part is understanding filter context, which explains why the same measure gives different results depending on where you place it. With a good data model, DAX becomes much more manageable.