Using anomaly detection
One of the great ways to find outliers is using the built-in Find anomalies tool in the Analysis pane of a visualization in Power BI. This feature of Power BI uses Artificial Intelligence (AI) capabilities to identify which data points are most different from other values in your data. The purpose of this feature is to help with analyzing the data so that action can be taken either against the data before business decisions are made or to help report consumers better understand the data and make informed business decisions.
To use this capability, you need to be using a supported visual (such as a line or scatter chart) and only one field on the y axis or values. This feature is being improved; so, at the time of writing, this was the requirement.

Figure 13.2 – Automated anomaly detection is built into some visualizations
Once this has been enabled, you can see the outliers identified with a gray dot. These are anomalies that have been identified. You can also set the sensitivity level, and even a categorical field to use as the basis for anomaly detection. With this capability, it is easy to quickly identify data values that are out of place or potentially outliers.
Now that we have an understanding of outliers, let’s take a look at how we can analyze time series data in the next section.
Conducting time series analysis
Time series analysis involves extracting meaningful data and identifying trends by analyzing your data in time order. You can use this time-ordered data to make predictions about the future, forecasting future needs or desired actions.
Visuals such as line charts can be used alongside the forecasting pane to show how your time series data is predicted to change. Many times, time series analysis involves using plots such as Gantt charts, which can be helpful for scenarios such as project planning or monitoring stock market data. By looking at your data using these tools, you can best identify when events have influenced change in your business. Any events that impact your business should be represented in the data and thus be identifiable when looking at the time series data.
In Figure 13.3, we can see the phases of a project as it moves along the axis of time. Before Feb 07, we can see that the phase was Scope, while the Analysis phase started from Feb 07 and has an end date before Feb 14. We can monitor the process of each phase as some will overlap and using this time series method, we can monitor the change over time and observe how events will impact the data.

Figure 13.3 – Gantt charts are a staple of modern business, and a classic example of time series reporting
Another way to add time series information to your reports is by using Play Axis. The scatter chart in Figure 13.1 has a play button in the lower left-hand corner. Clicking on it will have the chart “play” through the dates; you can watch your data change over time.
If you have a whole page of visuals you want to “play” over time, there is a custom visual called Play Axis that may help here, and it is available in Microsoft AppSource.

Figure 13.4 – Adding the Play Axis visualization to a report page
The Play Axis slicer can make a whole page of your report display your data ordered by date.