This page has 0 threads | Add post

Blogs

Finding our blogs useful? Please consider helping us to share more!

Power BI July 2022 Update - Error Bars
You can now display error bars in charts, as well as formatting them and adding error labels. We've been creating our idiosyncratic monthly blogs on Power BI updates since November 2016, and also deliver online and classroom Power BI courses.

Posted by Andy Brown on 25 July 2022

In this blog

Adding Error Bars to a Chart
Types of Error Bar (and Calculation Method)
Formatting an Error Bar
Error Labels

Power BI July 2022 Update - Error Bars

There are lots of teases in this month's update (offering filled maps for the Azure map visual is particularly cruel when you consider how long it's been languishing in limbo), but the only major release change is error bars.

These error bars have error labels and markers.

This blog shows how Power BI calculates error bars (I think), and how you can customise how they're displayed.

One other change: Power BI nerds (can't think Wise Owl have any of them) will be pleased to see that the Excel NetworkDays function has also now been carried over into DAX.

Adding Error Bars to a Chart

Having created a chart, you can enable its error bars like this:

Click on this option on the Analytics tab for the visual (the one that looks like an angry Pac-Man) and set the Enabled property of the Error bars card.

Types of Error Bar (and Calculation Method)

There are 4 types of error bar:

You can show error bars by percentage, percentile, standard deviation or by the value of some field.

Here's what these mean:

Option	What it means
By field	You can use the value of some other field for the upper and lower bound of each error bar (presumably this other field would be based on a DAX measure).
By percentage	You can specify that you want the bars to extend above and below the column or shape they pertain to by X or Y percentage of the number represented by the column or shape.
By percentile	To be honest I'm still trying to work out what this means statistically! The results seem odd no matter which values you choose.
By standard deviation	The usual statistical treatment is to show error bars within two standard deviations of the mean

Consider this chart:

For this chart we've set the error bars to be 2 standard deviations high.

Clicking on the left-hand data point gives this:

You can show the data for any point in a chart like this.

This gives this data:

The underlying data for the 2016 data point. These points return a standard deviation of 1.48, twice which is 2.96 - which rounds to 3 (hence the error bar).

Formatting an Error Bar

You can format an error bar within inches of its life:

Some of the things you can change about an error bar.

Here are the settings used for our error bars:

The format properties for the error bars shown above.

Error Labels

In addition to error bars you can also (as we've done) show error labels, which are similar to data labels for a chart:

We've chosen to show the labels as relative (numeric), so they show the offset from the chart's data point as plus/minus numbers.

If you care about error bars, you'll be delighted to see that they have been comprehensively incorporated into Power BI. They do provide a quick first-glance estimate for the reliability of any data underlying a chart.

Some other pages relevant to the above blog include:

This blog has 0 threads

Add a new post