A Comprehensive Guide to Power BI Error Bars

Introduction

Error bars are essential for visualizing data variability and uncertainty, making them invaluable for more accurate and meaningful analysis. In this guide, we’ll show you how to add and hack error bars in Power BI, helping you enhance your visuals and present more reliable insights.

Error bars can be applied to both Line and Bar charts, though they are not compatible with stacked charts. In this guide, I will explain the fundamentals of how error bars work and show you how to manipulate them to create various types of charts. By the end, you’ll not only understand the basics but also be inspired to use error bars creatively, expanding your visualizations with greater flexibility and imagination.

To access error bars: 1. Select the Visual > 2. Go to Analytics Tab > 3. Select Error Bar Option

Parameters of Error Bars

Error bars offer six key options:

1. Series: Choose the measure on which the error bars will be applied.

2. Options: Define the values that represent the upper and lower limits of your error bars.

3. Bar Settings: Customize the appearance of the error bars themselves.

4. Marker Settings: Adjust the markers that indicate data points on your chart.

5. Labels: Control the visibility and formatting of labels associated with your error bars.

6. Tooltip: Customize the information displayed when hovering over the error bars.

1. Series

For the series, I have selected Sales LY.

2. Options

In the Options section, you can define the Upper and Lower Bound values, which set the starting and ending points of the error bars. The specific values for these bounds are not critical unless you're using an arrow marker, where switching the values will reverse the direction of the arrows. Make sure to select the "Enable" option to display the error bars.

To base the bars on the upper and lower bound values, set the Type to "By Field" and Relationship to Measure to "Absolute," meaning the bounds are defined by actual values. Alternatively, choosing "Relative" under Relationship to Measure indicates that the fields represent the difference between the bounds and the measure.

For example, if the measure is 60, with an upper bound of 90 and a lower bound of 45, the absolute values would be 90 (upper) and 45 (lower). In relative terms, the upper bound would be 30 (90 - 60), and the lower bound would be -15 (45 - 60).

3. Bar

The bar settings allow you to adjust the width and borders to modify the thickness of the error bars. While the minimum width is 1, you can set the border size to 0 to make it invisible. There's also an option to match the bar color to the selected series.

4. Marker

For markers, you can choose the size and type, but not the color—it will automatically match the bar color. Wondering how to apply conditional formatting to markers? I’ll cover that in the Tips and Tricks section below.

5. Error Labels

In the Error Labels section, you can customize which values are displayed using the Label Format option. By default, the actual upper and lower bound values are shown, but you can also display the value or percentage of these bounds relative to the series.

6. Tooltip

Similar to error labels, you can display these values in the Tooltip by enabling the option. The same Label Format settings apply, allowing you to show the actual upper and lower bound values or their value/percentage relative to the series.

These are the basic options, but by combining and manipulating them, you can achieve a lot more.

Tip and Tricks

1. If you set the same value for both the upper and lower bounds, the bar size will be zero, leaving only the markers visible.
2. You don’t always need to provide both upper and lower bound values. You can input just one and use the Relative option under Relationship to Measure to calculate the difference from the series value.
3. Additionally, you can set the bar’s transparency to 100% in the color options, making the bar invisible.
4. It's also possible to overlap multiple error bars. For example, you can create one error bar for positive values and another for negative values. Since conditional formatting isn’t available for markers, you can create two error bars with different marker colors. By overlapping the actual bars from the visual, the markers will align perfectly. This approach will become clearer in the examples below.
5. You can even create dummy measures to overlap multiple error bars and markers for more complex visualizations.

Examples

To create this Enclosed Dot Plot, we need three components: the grey rounded bar in the background, the black marker for Previous Year, and the brown marker for Current Year.

To achieve these three components, we’ll need three measures. The actual values of these measures don’t matter, as only the error bar values are important. Therefore, we can use dummy measures with zero values, such as Sales CY or Sales LY. Since we will set the original bar to be transparent, these measures won’t be visible, allowing us to focus only on the markers and error bars.

Remember, any of these components can be built on any series since we are using Absolute values, not Relative.

For the grey bar, I used Sales CY and Sales LY, setting the bar width to 10 and the marker size to 7.2 so that the markers blend with the bar, creating a cohesive look. I then created a black dot for the Last Year Sales and a brown dot for the Current Year Sales, with the marker size set to 6 to ensure they are enclosed within the grey bar. Additionally, I turned off error labels for these markers to prevent overlap with the grey bar's error labels.

Finally, for the visual bars, I set the transparency to 100%. In the layout settings, I enabled overlap and set the space between series to 100%.

Similar techniques can be applied to manipulate the error bars and create the visualizations shown above.

You can find the link to the file here, where you can explore all the charts shown in the picture above.

I hope this helps you understand how it works. If you have any questions or need assistance with a specific scenario, please leave a comment. I’d be happy to help!