Free data visualization with Microsoft Power BI: Your step-by-step guide

We'll show you how to analyze a file with more than two million records of U.S. airline flight delays in this hands-on tutorial with video.

Executive Editor, Data & Analytics, Computerworld |

Microsoft has jumped into the free, self-service data analysis space with Power BI.

Power BI offers basic data wrangling capabilities similar to Excel's Power Query. It also lets you create interactive visualizations, reports and dashboards with a few clicks or drag-and-drops; type natural-language questions about your data on a dashboard; and handle files that are too large for Excel.

It can work with dozens of data types -- not only Excel, Access and CSV files, but also Salesforce, Google Analytics, MailChimp, GitHub, QuickBooks Online and dozens of others. And, it will run R scripts -- meaning that any data you can pull in and massage via R you can import into Power BI.

A guide to Power BI

In this article, I've put together a step-by-step guide to starting with Power BI, along with a number of other resources to help you along:

What is Power BI exactly?

Power BI includes both a downloadable desktop program and a cloud service, each of which offers different but overlapping capabilities. Data wrangling is desktop-only; visualizations and reports can be created in either; dashboards and report sharing are cloud-only. In addition, there are mobile apps for iOS, Android and Windows that let you view your Power BI or SQL Server Reporting Services (SSRS) reports and dashboards.

At least for now, you can take advantage of most Power BI capabilities without paying -- although Microsoft is clearly betting that you'll like the basic cloud service enough to spring for a $9.99/month paid account. Chief benefits of the paid account are increased data storage (10GB vs. 1GB), more timely automated data refreshes, the ability to create enterprise "content packs" and higher streaming capacity.

Be advised, though, that Microsoft wants a business email address when you sign up for Power BI cloud service -- while it can't screen out all non-commercial addresses, it won't accept known free consumer addresses like Gmail.com. Accounts from .gov and .mil addresses aren't supported for direct sign-up at powerbi.com either, although addresses at .edu and .org are.

And if you'd like to use any of Power BI's free mobile apps, you'll need a Power BI cloud account or access to your organization's SQL Server.

On the other hand, Power BI Desktop (at least for now) is not only free but doesn't require an account, an email address or a credit card -- just a Windows PC.

If you'd like to learn how to use this new, still-evolving tool to create reports and dashboards, read on.

Importing data into Power BI

Power BI Desktop is the better place to begin, unless you're sure that your data is already in the format you need for visualization. (Which may be the case if, like me, you prefer to do your data wrangling with a scripting language like R or Python.)

If you're used to Excel, you might think that selecting File > Open is the way to start analyzing your data in Power BI. But you'd be wrong -- File > Open is only for an already existing Power BI project.

Instead, to import new data, click the Get Data button on the Home tab, choose your data source type and click Connect.

001 powerbi get csv — To load a file, go to Get Data and select your data type.

This will bring up a familiar Windows file-selection dialog. Choose your file and you'll see a preview of your data. If it looks okay and there's nothing more you want to do to the data before starting to graph and chart, hit Load. Otherwise, click Edit, which brings up the Power BI Query Editor.

In this article, I'm going to use monthly files of airline flight-delay information from last summer that I downloaded from the Federal Aviation Administration (FAA) website. I know -- especially where airlines are concerned, past performance is no guarantee of future results. But if you're going to book a flight this summer, it might be fun (if not necessarily predictive) to answer questions such as: Which airlines had the best and worst delays last summer? Are there any specific flights that do especially poorly or well? These Power BI charts can help you easily answer these questions.

If you want to follow along, you can download your own data files from the Department of Transportation website. Or if you want, you can download the same files I'm using here -- the file download is available to all members of the Computerworld Insider program; registration is free, so if you're not already an Insider, it's easy enough to sign up. Files include data for domestic flights in the U.S. by month (so if you want to check flights to Paris, this won't help). There are separate files for June, July, August and September.

Start by loading in the June file (2015_06_ONTIME.csv): Go to Get Data > CSV in Power BI. Select and open your file, and you'll see a preview of your data. Then click Edit (not Load) to bring up the Query Editor. Now we can do some data wrangling.

One thing that can be useful to check at this point is whether number columns are loading in as numeric (aligned to the right) or text (aligned to the left). In other words, if you see numbers that are flush left in your data preview, they're not importing correctly -- which is one reason to choose Edit and bring up the Query Editor window even if you don't think you need to make changes in your data's structure.

002 powerbi datapreview — Make sure that the number columns are loading in as numeric -- in other words, aligned to the right.

Once in the Query Editor, you can right-click on a column header and select "Change Type" in order to manually select a data type such as whole number, decimal number, date, date/time, etc. But there's plenty more we can do with this data besides checking column types.

003 powerbi queryeditor — Once in the Query Editor, you can start tweaking your data.

Note: If you're not interested in data wrangling and want to get started with charts and graphs, load the summer15delays.csv file instead and skip ahead to the Easy visualizations section -- but do make sure that the flight number is changed from numeric to text when you import the file.

[Continues on next page]

Power BI Desktop vs cloud service

It can be confusing for beginners to understand when to use Power BI's desktop software and when to use the Power BI cloud service, since they do some but not all of the same things.

To begin with, the desktop application runs only on Windows. The cloud service is multi-platform, running on Microsoft Edge, Internet Explorer, Chrome, Safari for Mac and Firefox. In addition, mobile apps are available for iOS, Android and Windows 10 for viewing your reports and dashboards.

The desktop app is designed for authoring. It's where you can do data modeling, merging different data sources and so on as well as visualization.

The Power BI cloud service is for sharing and creating dashboards. It's also where you can access features like natural-language questions ("Hey Cortana...."). Cortana natural-language technology is getting baked into Power BI cloud. You can build visualizations and reports in the service, too, but you can wrangle data only in the desktop app.

Here's a bit more detail:

If you need to massage your data for analysis -- combine data sources, add or delete columns, or reshape tables -- you'll want to use the desktop software, since it's the only option for data modeling.
Both the desktop software and the cloud service do basic data visualizations and reports. I asked Microsoft about the difference; a rep told me via email in May: "For simple data visualization and exploration Power BI Desktop and the Power BI Service offer similar capabilities, but you'll encounter some differences. For example, R visuals aren't available in the service yet, but are planned for later this summer." (Update: R visuals were added to the cloud service in early July; a paid Pro account is needed to use them.)
To create dashboards -- visual summaries of data from one or more different sources -- and ask natural-language questions about data on those dashboards, you need the Power BI cloud service.
To get some rudimentary automated analysis (Quick Insights) from your data, you need the cloud service.
To share your analysis with others, you need the cloud service.
Some online data sources, such as Adobe Analytics, require using the cloud service.

You can start off working in Power BI desktop if you need its data-wrangling capabilities and then publish your data or report to your cloud account to create dashboards and share your work.

1 2 3 4 Page 1 Next

Page 1 of 4

It’s time to break the ChatGPT habit