UPDATE: April 2, 2021 – this post was published as satire on April Fool’s Day and is not factual

In a surprise move, Microsoft has decided to bring data visualization to Power BI. Long considered a data modeling tool used in support of exporting for analysis in Microsoft Excel, the ability to visualize data effectively has eluded Power BI users until now.

Previously unknown to Power BI’s 17 billion active users, data analyst Martin Vanderpfluegger discovered visualizations had come to Power BI while watching a Saturday morning Guy in a Cube livestream. “Patrick diverted from all the talk about Power Query and DAX. He clicked into this area he called the report canvas, and suddenly I saw this set of icons in a new section called Visualizations. At first, I thought maybe Patrick was accidentally using a preview version since he works for Microsoft, but then I opened my own Power BI Desktop and discovered this easter egg was real.”

Vanderpfluegger immediately posted his find on Reddit and Twitter. “One of my seven followers retweeted it, and it went viral from there. It got four, maybe five, Likes after a few hours. Reddit was even better. I’m still new over there, but it got hyphen 17 votes, and one person even replied with ‘WTAF?’, which I think might be a newer DAX function, and a few smiling emoji. It’s just great to see this level of community engagement.”

Almost everyone from Microsoft’s Power BI product team has deferred comments until after the official blog post arrives. When pressed, however, visualizations were reluctantly confirmed by one employee. “Who hoo hoo hoo”, said Senior Program Manager Henry the Owl about the development.

“I really think this is a game changer”, continued Vanderpfluegger. “The charting in Excel has been so static, and these clicky clicky visuals are cool. I had also been stuck in my org’s Excel to Power BI to Excel cycle for so many years. I’d build out these wicked star schemas in Power BI and copy and paste my dimension and fact tables into Excel, hook them up with VLOOKUP, and then visualize my data in Excel pivot tables. Power BI can now save me a few steps plus give me a variety of bars, lines, and pies. The matrix looks cool too. I can’t wait until they add more formatting options. I want to change my bar colors.”

Henry the Owl expects data visualization in Power BI to exit preview and become generally available before 2027.

The post Microsoft Power BI Shifts Focus to Include Data Visualization appeared first on DataVeld.

The default Power BI report layout uses a landscape orientation. While it’s not common to see report pages that use portrait mode, Prathy Kamasani (blog/twitter) has a notable example in Power BI here. Others have tried it as well. It’s used occasionally in other data communities or data-driven journalism. Overall, if you want to guide someone through your analysis by having them scroll like a web page, it may be an option to try.

So what happens if I create a report with a vertical layout in Power BI?

Not much. Except you might get questions. Some may begin with “Why did you do use a vertical layout?” You might be okay with a “tall” page until you hear “Why does this Power BI report take a long time to load?”, at which point, you’ve gone too far…

Personally, I do not see a vertical layout as a good option in most circumstances–with one critical exception. I think it’s at least worth a try if you have the right combination of:
1) Publish to Web or Embedded report with a small number of visuals and that has width constraints on the web page into which you’re embedding (blog or news page with sidebar, for example).
2) Viewers who are not familiar with Power BI’s interactivity and navigation options (typically external, public viewers)

With internal reporting or for users already familiar with Power BI’s interactivity like drill through, cross filtering, advanced page navigation and more; use a more traditional layout with which people are familiar.

UPDATE: Embedded “in the wild” Example

After initially posting this, Nick Nigro shared an example with me where it’s used and embedded outside of Power BI’s Data Stories Gallery. Let us know what you think, and thanks Nick!
https://securingdemocracy.gmfus.org/hamilton-dashboard/

Format: Page Size

Under the Page Size format area, Power BI has different Type options. While the default uses a 16:9 aspect ratio, choosing the Custom type and manually adjusting the Width and Height allows you to define any size and resulting orientation you would want.

For example, a simple vertical layout might involve swapping the default width and height so that 720 pixels are now used for the width and 1280 pixels are used for the height. You could define height at any (well-performing, reasonable) value based on the needs of your report though.

Don’t Forget “Fit To Width”

The default Page View option in Power BI is Fit to Page, but when using a vertical layout, the entire page appears on screen. This default makes it more difficult to design. Instead, if you attempt a vertical layout, go to the View tab in the ribbon, select Page View, and then choose Fit to Width. Actual Size might be reasonable too depending on your dimensions.

This option stretches the report canvas to take the available width, and you scroll up and down to add visuals along your vertical layout.

Pros

• Mimics the natural scrolling behavior of someone browsing a website.
• Gives the viewer more control over their navigation.
• More content fits on a single page (although see a Con below if you add to many visuals to a page).
• Good top to bottom story format.
• Could help with more easily viewing content if public Publish to Web reports are embedded in sites with reduced width (such as a news page or blog with a sidebar) or if the audience will largely be viewing the Publish to Web report on mobile.

Cons

• Poor format for reports that are highly interactive, rely on a lot of cross filtering for analysis, bookmarks, etc.
• Scrolling could be confusing to viewers at first since this page orientation is not commonly used in Power BI.
• May be difficult to find specific visuals compared to standard page navigation where entire pages appear on screen.
• Viewer will not know if the report page is 1,000 pixels or 20,000 pixels (how tall is tall?).
• Since a single page can be as tall as you want, performance may suffer as you add more and more visuals to the page.
• Export to PDF or PowerPoint does not maintain the Fit to Width setting, so a single page in Power BI is still a single page when exporting.

Can you think of any more Pros or Cons of using a vertical report layout in Power BI? I’d appreciate your comments below.

The post Power BI: What Happens if I Create a Report with a Vertical Layout (Portrait Mode)? appeared first on DataVeld.

ORIGINAL POST from December 2019:
With the December Power BI Desktop update, users are introduced to a new color picker when customizing theme colors. The new color picker is much more versatile than the original when selecting colors.

Power BI’s new custom color picker:

Power BI’s original custom color picker:

The new color picker allows colors in RGB format in addition to the hex color format that Power BI has used exclusively until now.

The new one also easily allows users to choose from a wider selection of shades and tones. This builds upon the simpler selection of hues and tints in the original.

Oversimplifying color definitions:

• Hue – the dominant color along the spectrum of red to green to blue back to red (this is usually visualized as the “color wheel“)
• Tint – color when a range of white combines with hue
• Shade – color when a range of black combines with hue
• Tone – color when a range of grey combines with hue

The new color picker represents Hue along the bottom bar. If this were shown on a color wheel, it would be a 360 degree circle moving clockwise from red to green to blue to red. In the case of Power BI, you can adjust hue by clicking in the bar and dragging the selection or by clicking and hitting the left or right arrows keys on the keyboard. In fact, if you use the arrow keys, you’ll see that you increase or decrease hue values by the equivalent of a degree on a circle. Give it a try–you can start on the left and hit the right arrow 360 times to get all the way to the right (don’t actually try this 360 times…).

Tint, Shade, and Tone are achieved by clicking and dragging the small circle in the square above the hue bar. You can also manually specify the hex or RGB values. As a result, the new color picker better visualizes the full range of tints, tones, and shades for a specific color.

In contrast, the original color picker:

• does not allow for wider tone or shade selection
• shows only hex colors and not RGB
• does not change the hex value when hovering over the image (this makes precision selections impossible through a visual hover and click)

The post A Primer for the Microsoft Power BI Color Picker appeared first on DataVeld.

Let’s take a look at how to build it.

Data

Obtain a dataset that has coordinates in three separate columns. In my case, I needed to add an Index column so that I could break out all of the points into unique rows once added to a visual in Power BI. Your data might look something like this:

What-If Parameters

The dynamic rotation relies on “what-if” parameters. Add three “what-if” parameters corresponding to the three axes. In my case, I used possible rotation values of -90 to 90 degrees, but you could use anything from 0 to 360, -180 to 180, etc. Be sure to leave the option to “Add slicer to this page” checked. Having the slicers is what enables users to dynamically rotate your object.

X and Y Measures

The projection from three coordinates to a 2D plane is achieved by adding the following two measures. Be sure to adjust the column references and what-if parameter names at the top to correspond to your own data.

Here’s my “Ortho x” measure. The initial six bold values are what you’d need to adjust to your own data and parameter names.

Ortho x = 
VAR x =
    MIN ( Coordinates[x] )
VAR y =
    MIN ( Coordinates[y] )
VAR z =
    MIN ( Coordinates[z] )
VAR xDegrees =
    RADIANS ( 'X Rotation'[X Rotation Value] )
VAR yDegrees =
    RADIANS ( 'Y Rotation'[Y Rotation Value] )
VAR zDegrees =
    RADIANS ( 'Z Rotation'[Z Rotation Value] )
VAR x0 = x
VAR y0 =
    y * COS ( xDegrees )
        + z * SIN ( xDegrees )
VAR z0 =
    z * COS ( xDegrees )
      - y * SIN ( xDegrees )
VAR x1 =
    x0 * COS ( yDegrees )
        - z0 * SIN ( yDegrees )
VAR y1 = y0
VAR finalX =
    x1 * COS ( zDegrees )
        + y1 * SIN ( zDegrees )
RETURN
    finalX

Here’s my “Ortho y” measure. Again, the initial six bold values are what you’d need to adjust to your own data and parameter names.

Ortho y = 
VAR x =
    MIN ( Coordinates[x] )
VAR y =
    MIN ( Coordinates[y] )
VAR z =
    MIN ( Coordinates[z] )
VAR xDegrees =
    RADIANS ( 'X Rotation'[X Rotation Value] )
VAR yDegrees =
    RADIANS ( 'Y Rotation'[Y Rotation Value] )
VAR zDegrees =
    RADIANS ( 'Z Rotation'[Z Rotation Value] )
VAR x0 = x
VAR y0 =
    y * COS ( xDegrees )
        + z * SIN ( xDegrees )
VAR z0 =
    z * COS ( xDegrees )
        - y * SIN ( xDegrees )
VAR x1 =
    x0 * COS ( yDegrees )
        - z0 * SIN ( yDegrees )
VAR y1 = y0
VAR finalY =
    y1 * COS ( zDegrees )
        - x1 * SIN ( zDegrees )
RETURN
    finalY

Scatter Chart

Once you have the two new X and Y measures, add a scatter chart visual to the report. Drag your X measure to the X Axis, Y measure to the Y Axis, and whatever you use for a unique index to Detail.

Conditional Formatting

To give a better visual appearance to depth along the z axis, add conditional formatting to the scatter chart. Go to Format and expand Data Colors. Hover over the area and click on the […] that appears, and then select Conditional Formatting.

Since my depth column is simply named z, I chose to base the color scale on “sum of z”. Choose your own field as appropriate. Select a color scale that you would like and any other options before clicking OK.

Finally, I adjusted some other format options like Shape to reduce the point size. With everything built, I now had a “face” in my data that took the 3D coordinates from my original dataset and projected onto the 2D Power BI scatter chart.

Try it!

Here’s a live version of the report to try out, and you can download the sample PBIX file from the Data Stories Gallery.

The post 3D Effect with Orthographic Projection in Power BI appeared first on DataVeld.

While there is more to the report than what the videos cover, the primary areas of interest include:

• Report page tooltip – additional filtered visuals appear in the tooltip when hovering over an element
• Bookmarks and Buttons – navigation is controlled via Bookmarks, which are called by the Trends, Compare, and Back buttons
• Conditional formatting with selectable measure values – the periodic table uses a scatter plot with conditional formatting, but there are several measures at work and not a single one
• Dynamic axes – a scatterplot has dynamic values for both the X and Y axis rather than a static selection

The post Videos for the Power BI Periodic Table of the Elements appeared first on DataVeld.

The Power BI Blogmap is based on the Google News “Newsmap” treemap from about 15 years ago. The idea is a simple adaptation:

• There is a weight assigned to the recency of posts.
• Individual blogs are different colored categories.
• The categories display the last ten posts for each blog.
• Posts published in the last 24 hours have the highest weight and will have the largest individual rectangles.
• Overall size and position of the individual blogs within the treemap are driven by the overall recency of the last ten posts.
• The blogmap refreshes every four hours.

Click the image below to visit https://powerbi.world/ to check it out!

The post Power BI Blogmap at PowerBI.World appeared first on DataVeld.

There are a few pieces that need to be assembled to enable data sonification that works in both Power BI Desktop and Service. While, audio has been used in the Power BI Service for awhile, Power BI Desktop has been silent until now. This post attempts to show how to produce audio tones in Power BI for greater accessibility. It also demonstrates how to blend data with a standard range of audio pitches.

An embedded sample appears below. Note that audio in Publish to Web reports embedded into a page may not work, so it would be best to use this direct link to try the sample: https://app.powerbi.com/view?r=eyJrIjoiMGFkYjVjNjYtOWZlZi00Y2E5LTk2YWQtYmZlMzk1MWRkY2Y5IiwidCI6ImFjYzhhYWE1LWYxOTEtNDgyZi05MjFiLWNmNmMzM2E1ODgzMiIsImMiOjF9

In addition, most browsers now limit Autoplay for sound, but this setting can be allowed in both Chrome and Firefox using an icon near the address bar. Try changing the setting to allow sound and then refresh the page.

Audio Frequencies

In this sample, low pitch corresponds to low data values while higher pitch corresponds to higher values. The sample represents a range of frequencies from 130.81Hz to 2093Hz, which aligns with musical notes C3 to C7.

I generated a series of Wav files at the appropriate tones and stored them in a local folder for later import into Power BI. I’ve placed my core set of files from this solution into a GitHub repository here: https://github.com/deldersveld/datasonification

HTMLViewer with HTML5 Audio

As with past audio solutions in Power BI, attempting data sonification in Power BI relies upon the use of the HTMLViewer custom visual along with the <audio> HTML tag. The format is fairly basic and only requires a src and autoplay specified.

<audio autoplay src='[...]'></audio>

Autoplay is important in this case because users would otherwise need to click “play” on visible audio controls for every data point. Using autoplay may have an impact in Publish to Web reports on Google Chrome, but it seems to work fine with reports published to the Service otherwise in Chrome. Firefox and the Chromium version of Microsoft Edge worked fine for me when tested.

Imported Base64 Audio in Power BI

The method imports audio into the data model rather than relying on hosted files. This allows the audio to play in Power BI Desktop as well as Service. Desktop never traditionally played audio from hosted wav or mp3 files well.

The technique is the same one used for Base64 images. It relies on converting wav audio files to Base64 and prefixing the converted audio with a data URI.

data:audio/wav;base64,
data:audio/mp3;base64,

In the Power Query Editor, the key steps are to get data from a folder containing Wav files (or MP3) and then concatenate the HTML audio tag text around the Base64-converted binary.

let
    Source = Folder.Files("C:\Data\Wav"),
    #"Added Custom" = Table.AddColumn(Source, "HTMLAudioBase64", each "<audio autoplay src='data:audio/wav;base64," & Binary.ToText([Content],BinaryEncoding.Base64) & "'></audio>")
in
    #"Added Custom"

As with Base64 images, however, there is a tradeoff. The column containing audio cannot exceed the character limit of a Power BI text field, which is roughly 32,000 characters. This limitation means that imported audio is limited to short durations and small sample rates. You will not be playing music with this method. The sine tones used in my sample are only 8kHz and .25 seconds in duration.

My final imported table contains the audio column along with Frequency, a label for the associated note at that frequency, and an index column used for sorting and a table join that is independent of frequency.

Joining Another Table to an Audio Table

Once I have a table containing the audio, I join the audio to the data I want to sonify. I intentionally deviate from two “best practices”: the use of calculated columns and cross filter direction. Since this is a special case, there is no regret of course.

How can numeric data with a wide variation of values align with audio frequencies? The DAX formula for the calculated column maps data values proportionally to available audio values. Wrapping CEILING around the formula rounds the number to an integer so that it can be joined to the index column in the audio table.

Audio Join =
VAR __minAudio = MIN(Wav[Sort])
VAR __maxAudio = MAX(Wav[Sort])
RETURN CEILING(DIVIDE(Superstore[Sales],MAX(Superstore[Sales]) - MIN(Superstore[Sales])) * (__maxAudio - __minAudio),1)

Since this relies on the numeric measure field coded into the calculated column, it reduces flexibility. Visuals that rely on additional filtering in explicit measures will not “play” appropriately. Data sonification is indeed a special case. You need to sacrifice some interactivity by structuring the table to correspond to values in the final visual. For example, perhaps you have a transaction table with orders and sales at the grain of Order ID. To apply data sonification to a Sales by Order Date, you would need to create a summary table with SUM(Sales) grouped by Order Date.

The audio is on the 1 side of a 1:* relationship, and the audio column needs to filter along with the visual data. To facilitate the appropriate filtering to the correct frequency, the cross filter direction on the relationship needs to be Both.

Visualization Setup

With the data model in place, the next step is to obtain the HTML Viewer custom visual from the Marketplace and add it to the report canvas. Add the column that contains HTML audio to the visual.

The “Bottom 1” visual level filter on the HTML Viewer visual ends up being important. Power BI (really the HTML Viewer) plays all audio in the column whenever there is no selection. The visual level filter set to Bottom 1 prevents 30 tones from playing at once on report load or when the data tones finish playing. I chose Bottom 1 instead of Top 1 because the lowest pitch is less noticeable and more tolerable to my ears than the highest pitch.

Since I did not specify controls in the <audio> tag, nothing appears visually on the report. In general with audio, it would be a good practice to show, but I did not include it since the audio sequence is going to be controlled separately.

Obtain the Play Axis custom visual from the Marketplace and add it to the report canvas. Add the column that corresponds to the axis on the sonified visual. In my case, the main visual consists of Sales by Order Date, so the Play Axis visual needs Order Date.

Turn on the volume and hit play!

Sample PBIX

Download the PBIX file from GitHub or the Power BI Data Stories Gallery:

https://github.com/deldersveld/PowerBISamples/blob/master/Data%20Sonification.pbix

https://community.powerbi.com/t5/Data-Stories-Gallery/Data-Sonification/td-p/756842

The post Data Sonification in Power BI appeared first on DataVeld.

Try It!

What is unique for the user is not the way the primary visual is filtered by other visuals–it’s how the filters are presented using an accordion-like control.

Try it out live:

Filter Pane

If you use Power BI’s recently released Filter Pane, you should notice that filters on that pane behave in a similar manner. Accordions are not novel–they’re ubiquitous in many different types of applications. The approach that I took for the custom accordion filter would be like adding the ability to use any type of visual with Power BI’s Filter Pane.

Custom Accordion

The Power BI custom accordion relies on Bookmarks and Buttons as key elements. I’ve only created two categories in my accordion. I’ll be honest–it’s probably more work than it’s worth to keep track of different buttons due to positions as well as what’s visible or hidden for each bookmark. The thought of expanding to three categories is a bit daunting. Why is that?

Bookmarks

If you have worked with bookmarks in Power BI, you might already be able to guess what the bookmarked setup visibly looks like. For this sample, there are three bookmarks:

The Main bookmark has the accordion elements collapsed.

The State Expanded bookmark shows the State bar chart and moves the position of the Year button.

The Year Expanded button shows the Year bar chart.

Buttons

The bookmarks seem straightforward, right? It’s simply a matter of showing or hiding certain objects as with any bookmark toggle.

The need for several different button objects and positions is where a higher level of complexity comes in though.

There are actually two overlaid State buttons, not a single State button. That is because the Action setting of the button is not “bookmarkable”. A single button cannot currently launch two different bookmarks. Therefore, the main State Button has an action to view the State Expanded bookmark and the State Reset button has an action to view the Main bookmark. Depending on whether the user is in Main or State Expanded, only one of the state buttons will be visible at a time.

The Year button is more complex because of the position change with the State Expanded bookmark. The Year button shifts downward to allow room for the State bar chart. As a result, there are two buttons in the initial position (one for choosing the Year Expanded bookmark and one for resetting to the Main bookmark). There is an additional year button in the lower position for when the State Expanded bookmark is in use.

I hope you can appreciate why I only tried this with two accordion selections. Three options require four individual buttons for the third category due to having to respond to two possible position shifts, four selections require five buttons for the fourth category, etc. As you can see, expanding the possibilities could be difficult to manage fairly quickly.

This post introduces the concept and is ultimately not a step by step guide to a Power BI accordion control. Please feel free to download the PBIX from GitHub and explore if you’d like to try to recreate this for your own report.

The post Build a Basic Power BI Accordion Filter (or Don’t) appeared first on DataVeld.

Paginated reports have a number of built-in chart options including tablix (table, matrix), gauge, maps, sparklines, and numerous other chart types.

While there’s nothing directly akin to true custom visuals in Report Builder, paginated “custom visuals” could be crafted using a few different techniques. How can we create different chart types in Report Builder when they are not one of the default chart options?

As an initial example, let’s consider a waffle chart. This chart provides the proportion of a total by visually coloring a section of the grid to match the contribution to total. Most waffles consist of a 10×10 grid that shows % of total using color.

The following waffle example reads from left to right.

The following waffle example reads from bottom to top.

There are at least three ways to accomplish a waffle chart like the above examples in Report Builder:

1) Tablix & Code

One method is to use a tablix and some VB code. Jason Thomas outlined this years ago. While this works for a waffle chart and perhaps other grid-style visuals, it does not have much versatility beyond the constraints of rows and columns in the tablix.

2) Shapefile

If you want more versatility, you can create a shapefile using a tool like ArcGIS or QGIS and add it to a Map layer. You can create any kind of design and go beyond the grid limitation of the tablix. The downside of a shapefile is that if you want to adapt it in different ways, you would need to create multiple shapefiles and somehow switch between them. A shapefile is static.

3) SQL Server Spatial Query

Another way to create “custom visuals” in Report Builder is to use either SQL geometry or geography data types in a spatial query and then add that field to a Map layer. The primary downside is that you need to use SQL Server to run everything. If you can get beyond that limitation, however, it opens up a world of opportunity.

How’s that? Isn’t SQL geography and geometry stored in a database, so it needs to be predefined just like the shapefile would? Typically, a SQL spatial query is. The data is expressed in a spatial dataset and then joined to a separate analytical dataset.

Why not create your own spatial dataset at runtime dynamically based on your analytical dataset?

SQL’s geography and geometry can be expressed in well-known text (WKT) format. Points, Linestrings, Polygons, and Multipolygons can all be stated in an easy to read and parse form. With this in mind, you don’t have to start with existing shapes, lines, or points. You can calculate your own coordinates and build your own geographic/geometric features dynamically in a SQL query. This allows you to generate visuals that are data-driven rather than rely on predetermined shapefiles or SQL spatial data.

Getting Started

The waffle chart is a good starting example because we only need four points to draw a square polygon. Squares are also easier geometry than many types of more complex shapes to help introduce the concept of dynamically creating a spatial dataset. For simplicity, pretend we wanted to dynamically simplify the 10×10 grid to 2×2 grid. In the n=4 square sample grid below, we would need to calculate the four coordinates that define the square as a polygon shape. Assuming that we’re working on a 2D plane that spans from (0,0) to (1,1), we need to determine the x and y coordinates for each square in relation to the entire grid.

While this 2×2 grid is highly inaccurate compared to the 10×10 grid, it hopefully gets the concept across.

How does dynamically generating the coordinates work? In the upcoming Part 2 of this topic, I’ll step through the basics of how I use recursive common table expressions in SQL to help generate coordinates dynamically.

Be careful though. Dynamically creating SQL Geometry or Geography is the Paginated Report Builder equivalent of creating dynamic SVG image measures in Power BI Desktop. No one asked for it, few have a use for it, it’s tough to comprehend, but it’s kind of a fun technique.

The post Paginated “Custom Visuals” in Power BI Report Builder: Part 1 appeared first on DataVeld.

Why do people use funnels?

It is well-established tradition that sales reports include visuals where an axis (usually the Y axis) has bars that do not align at zero. Instead, sales “pipelines” and other entities that narrow down from 100% should be visualized with staggered bars reminiscent of a physical funnel.

Why add bars inside the funnel to form a FUBAR chart?

Bar charts add the appearance of accuracy and professionalism to any visual. People trust bar charts. By adding a series of bar charts inside a funnel, Power BI is now able to satisfy both data visualization purists as well as people who require the visual appeal of the funnel.

How do you read a FUBAR chart?

The funnel part of the FUBAR chart represents the total. Each bar in the bar chart shows the contribution to the total. It’s that simple.

With the FUBAR chart for Power BI, now you too can share this advancement in data visualization with your organization. Add a comment below to get a link to the PBIVIZ and sample PBIX file.

****
UPDATED 4/3/2019: Please note the date of post as April Fool’s Day. The FUBAR visual does not exist.

The post Introducing the FUBAR Chart for Power BI appeared first on DataVeld.