Power BI is Business Intelligence software from Microsoft. According to Wikipedia, Business Intelligence “comprises the strategies and technologies used by enterprises for the data analysis of business information”. At the March general meeting, Sean Henderson gave us a very clear and knowledgeable introduction to the latest Business Intelligence software, focusing on Power BI.
Sean contrasted Business Intelligence software with report generators, describing BI as “Reporting Plus”. An audience member asked: “Is Power BI a report generator, like e.g. Crystal Reports?” Answer: Power BI is more than reporting, it’s an analysis tool.
Some of the typical features of BI software are:
- an ability to use multiple data sources;
- advanced charting…many different types;
- interactive real-time filtering…not paper-based…interactivity is key;
- advanced BI – some software understands spoken-word queries;
- prescriptive nature – not only historical – predictive analytics (this is the hardest part to put together).
Power BI is mostly a service – a lot of its functioning takes place on cloud servers. However, there is a desktop software piece, which is used for design work and interactive data analysis. (Sean mentioned that Microsoft provides an on-premise server option but it is expensive.)
Power BI has some unique terminology:
- “Visualization” = chart
- “Slicers” = filters
- DAX language – similar to Excel formulas
- “Measure” = aggregated formula
A report is designed in the desktop software and published to the Power BI service. From there, pieces of various reports can be aggregated into a “Dashboard”.
A very prominent feature of Power BI is its data handling capabilities, which go way beyond those of many other types of software traditionally used for data analysis and decision support. Power BI can use many different data source types, e.g. websites (including social media sites, such as Facebook), databases, analysis servers, any ODBC or OLE sources, Excel files, text files, etc. During the demo portion of his presentation, Sean showed us the very long list of data sources that Power BI can work with. In addition to its versatility, Power BI can work with much larger volumes of data than tools like Excel and Access.
A non-commercial, i.e. a corporate, e-mail address is required to use Power BI. The usual per-seat cost of the product is about $10 per month, although a limited-capability version is free. Sean recommended Power BI over Qlik (a competing BI product), which offers a little more functionality but is much more expensive.
Our speaker was able to demonstrate the interactivity of a dashboard, in contrast to conventional reporting tools, accomplishing this in spite of the fact that the hospital’s internet connection was not working at the time. (A part of the display that included a map of Connecticut, which requires the internet connection, was blank.) He demonstrated the accessing and joining of data sources, and showed us how to create
an infographic.
Sean shared some of his personal preferences with respect to the design or “tone” of infographics, notably that the use of color is important. Graphics that employ only one or two colors, or sometimes none at all, look dull and uninteresting.
Sean illustrated the versatility of Power BI by demonstrating its use to generate a variety of visualizations, including waterfall charts (see https://en.wikipedia.org/wiki/Waterfall_chart) and tree maps (see https://en.wikipedia.org/wiki/Treemapping).
Sean likened Visualizations to Firefox plug-ins – many have been shared by the community of Power BI users, and are available online.
Power BI visualizations are very flexible – we saw how there can be lots of variation in how the data are displayed.
Sean explained how the steps used to prepare data for Power BI are stored and can be re-executed to refresh the data. Refreshes can be programmed to take place automatically at regular intervals: free users get daily refreshes while paid users can have hourly. In addition, Power BI provides some data streaming functions, which can update a dashboard in real-time.
Some other facts about Power BI:
- Embedding code can be generated to place a Power BI visualization into a web page.
- Power BI includes permissions functionality that can be used to control who will have access to reports and components of them.
- Updates are released monthly.
- “M” and “R” are other programming languages used with Power BI.
- “M”, the Power Query Formula Language, is a powerful mashup query language.
- “R” is a traditional programming language that can be used to perform many different tasks associated with Power BI.
A member of the audience asked whether Power Query and Power Pivot (Excel add-ins) are related to Power BI. Sean answered that these have the same underlying engine as Power BI, providing some Business Intelligence capabilities in Excel.
Sean concluded his presentation with a short Microsoft video about Power BI. There are lots of videos on YouTube that explain more details about how to use Power BI.