Analyze Your Data Set Using Machine Learning

Machine learning analyzes the data in your data set to provide insights that enable you to explain the various aspects of that data.

Use Machine Learning to Discover Data Insights

Machine learning analyzes the data to recognize the patterns and trends in your data set to provide visual insights and enhanced statistical analysis. You can subsequently use these visual insights and statistical analysis in your project visualization canvas to interpret the data in your data set.

Machine learning provides accurate, fast, and powerful data insights because it analyzes and processes technical and statistical complexity and the volume and variety of the date in your data set. Because of machine learning’s accuracy, speed, and scale, it’s cheaper and more powerful than the traditional method of analyzing data.

To discover data insights, you simply select an attribute in your data set. Machine learning provides you with narratives, visual insights, and statistical analyses such as charts. You can select specific charts and include them as visualizations in your project visualization canvas. You manage these visualizations as you do any other visualizations in your project. With machine learning, you don't have to waste time guessing and dropping random data elements on the canvas to create a visualization for data insight.

Before you start, install machine learning on the Windows or Mac machine where you installed Data Visualization Desktop. See How do I install Machine Learning for Desktop?

After you’ve installed machine learning, you can start uncovering insights in your data. See Add Data Insights to Visualizations.

Add Data Insights to Visualizations

You can select specific data insights charts provided by machine learning and add them directly as a visualization in your project’s visualization canvas.

Tutorial icon Tutorial

You must install the Data Visualization machine learning component to display the Explain option.

  1. Create or open a data visualization project. Confirm that you’re working in the Visualize canvas.
  2. In the Data Panel, right-click a data element (attribute or measure) and select Explain <Data Element> to display the Explain <Data Element> dialog tabs:
    • Basic Facts about <Data Element> - Shows the basic distribution of the data element (attribute or measure) values across the data set and its breakdown against each one of the measures in the data set.
    • Key Drivers of <Data Element> - Shows data elements (attributes or measures) that are more highly correlated to the outcome for the selected data. The charts showing the distribution of the selected attribute value across each of the correlated attributes values is displayed.
    • Segments that Explain <Data Element> - Shows the segments or group in the data set, after examining all the records, that can predict the value of the selected data element. You can select a particular segment or group and then continue to analyze it.
    • Anomalies of <Data Element> - Shows the group of anomalies or unusual values in the data set that you can relate to the selected data element (attribute or measure). You can review and select particular group of anomalies.
  3. Use the Explain dialog to help you configure your visualizations.
    • When you click a data element (attribute or measure), information for the selected data element is highlighted in the segments below.

    • You can select more than one data element (attribute or measure) at the same time to see results in the segments.

    • You can also sort how the information is displayed in the Segments (High to Low, or Low to High, group by Color, or sort by data element Value).

    • For each Segment in the decision tree, summary rules for the percentage of the data element and other metadata about the section are displayed. For example, a certain Segment might show that a particular percentage of the selected attribute (data element) belongs to a specific group like location, data point, another attribute, or measure. You can then select a specific group, like location, to analyze the selected attribute.

    • The Anomalies section finds data points that don't fit the expected pattern.

  4. Click the check mark when you hover the mouse pointer over any of the data insight charts to select a specific chart.
  5. Click Add Selected to add the charts you’ve selected as different visualizations in your project’s visualization canvas. You can manage data insight visualizations like any other visualizations you’ve manually created on the canvas.