Power BI

Power BI is Microsoft's business intelligence tool. You build reports in Power BI Desktop, a free Windows app, by pulling in data, shaping it, defining measures and laying out charts on report pages. You then publish to the Power BI service, the cloud side, where colleagues view and interact with them. If your company runs on Excel, Teams and SharePoint, Power BI is often already there and already paid for.

The free desktop app makes it easy to start. Sharing privately is where licenses come in, and that decision usually belongs to IT.

Who it suits

Power BI suits people inside organizations that use Microsoft tools: finance teams building monthly reports, sales operations tracking pipeline by region, HR teams reporting headcount. It suits analysts who are comfortable in Excel and want something that refreshes automatically instead of a spreadsheet they rebuild every month.

It suits freelancers and hobbyists less, because sharing outside an organization is awkward, and Mac users less still, since the desktop app is Windows only.

What it does well

The data side is the strength. Power Query, the same engine behind Excel's "Get Data," handles cleaning steps such as splitting columns, unpivoting, merging tables and fixing types, and it records each step so the next refresh repeats them. The data model lets you relate several tables (orders, customers, products, a date table) so one filter works across all of them.

Measures written in DAX, Power BI's formula language, handle the business logic: year-over-year growth, running totals, rolling 28-day averages. Once a measure is defined, every chart can use it.

Other things it does well:

• Scheduled refresh from databases, SharePoint lists, Excel files and many online services.
• Row-level security, so each regional manager sees only their region.
• Integration with Teams, SharePoint and Excel, where reports can be embedded or analyzed.
• A large library of visuals, plus custom visuals from a marketplace.

Where it gets in the way

DAX has a reputation for a reason. Simple sums are easy, but anything involving filter context (why does this total ignore my slicer?) takes real study. Expect to spend time learning it if you go beyond basic reports.

The default charts are serviceable but cluttered. Out of the box you get legends, gridlines, axis titles and borders on everything. Good reports come from removing most of that, sorting bars, and adding direct labels where possible; the labels guide applies directly.

Formatting is spread over many panes and toggles, which makes small design changes slow. And the licensing model, with per-user licenses, capacities and features that differ between them, confuses most people the first time.

Free options

Power BI Desktop is free and full featured for building reports on your own machine. You can save .pbix files and share the file itself with someone who also has Desktop.

Publishing for others is a different matter. Viewing reports in the Power BI service normally requires the author and viewers to have paid licenses, unless the organization has a capacity that covers free viewers. There is also a "publish to web" option that makes a report public to anyone with the link. It is meant for public data only, and many organizations switch it off.

How it compares

Against Tableau: both build interactive dashboards with linked filters. Power BI is stronger at data modeling and usually cheaper inside Microsoft-based companies. Tableau is more fluid for visual exploration and gives more freedom in how a chart looks. Choosing between them is usually decided by what the company already uses.

Against Flourish: Flourish is for published, often animated stories on the web, built from templates in a browser. Power BI is for internal reports that refresh from live data. They rarely compete for the same job.

A good first project

Take a sales or web traffic export you already have in Excel, ideally one you update by hand every month. Load it into Power BI Desktop through Power Query, do the cleaning there instead of in Excel, and add a proper date table.

Write three measures: a total, the same total for the previous year, and the percentage change. Build one page with a line chart of the monthly total, a sorted bar chart by category, and a slicer for the year. Then delete every legend, border and axis title you don't need. Next month, refresh it instead of rebuilding it, and you will see why people put up with DAX.