Visualize Subtab

About: This article describes the Visualize subtab in the Analysis tab.

Table of Contents

• Introduction
• Univariate Analysis
• Multivariate Analysis
• Chart Types
• Visualization Options Sidebar
  • Statistical Tests
• Video Tutorial
• Related Articles

Introduction

As the name suggests, this corner of Predict is dedicated to visually representing data records.  This can be an important part of the model exploration process. By using the Visualize subtab, users can observe detailed relationships between a stand alone variable or between multiple variables. There are many ways to focus in on key aspects of data using the Visualize subtab.

Note: Some additional display choices, such as what colors to use, are available under the Tools > Options menu.

Univariate Analysis

Select the Univariate Analysis analysis option to explore the distribution of a single variable in a chart. The variable to explore can be selected from the Variable drop down list. The distribution can be represented through the user's choice of a bar chart or a pie chart, which can be selected in the Chart Type dropdown. This is helpful when exploring the content of a single variable within your dataset.

Multivariate Analysis

Select the Multivariate Analysis option to analyze two or more variables simultaneously, illustrating how pairs of variables may be related.

Y Axis

When the Multivariate analysis option is selected, the Y Axis variable becomes available as a drop down containing all binary and continuous variables from the dataset. The values of the selected variable will be placed on the y-axis and its relationship to the x-axis variable will be graphically depicted. If the selected y-axis variable is binary, the y-axis will be a range from 0-1 depicting the average value of the y-axis variable for all categories of the x-axis variable. 

X Axis

When the Multivariate analysis option is selected, the X Axis variable becomes available as a drop down containing all variables in the dataset. The x-axis will display categories (for categorical or binary variables) or a range of values for continuous variables along the bottom of the chart. The graph represents the relationship between the y-axis and the x-axis variables.

View By

When the Multivariate analysis option is selected and both the X Axis and Y Axis fields are populated, the View By drop down becomes available. The variable selected from the drop down list will be placed on the X-axis as a sub-category with a legend for the subcategories to the right of the chart. 

Show Related Variables

The Show Related Variables option allows users to establish a p-value threshold for variables that will be visually represented. The result is a color coded visual that indicates the statistical relationship based upon the selected p-value and the number of records of that variable within the dataset. To view a record count of each variable, hover over the variable in the chart.

Chart Types

Bar Charts

Bar charts represent categorical data. One axis of the chart shows the specific categories being compared, and the other axis represents measured values. Bar charts are commonly represented in the "Univariate" view.

Pie Charts

Pie Charts are circular graphs that divide into slices to illustrate numerical proportions. Each slice is proportional to the total number of individual values in the variable.

Bubble Charts

A bubble chart is a chart type that displays three dimensions of data. Data is plotted in a disc shape and expresses the value of data in relation to size and color. In the example below, the size of the discs (degree major) relate to the number of students with that assigned major, and the color of the disc relates to the frequency that category has made contact with the academic advisor. 

Bubble charts are great for showing multiple dimensions of information in a simple, easy to understand visual. Bubble charts are only available in Predict under the "Multivariate" view. Remember to select a categorical variable for the x-axis to create a bubble chart.

Visualization Options Sidebar

This Visualization Options sidebar can be found on the left edge of the chart. This menu is available in both univariate and multivariate viewing modes. It provides an array of display options such as color, fonts, labeling, and data order. The options themselves change depending on which viewing mode the user has selected. This view may be "pinned" permanently open by clicking the pushpin icon in the upper right corner of the sidebar frame. Clicking the pushpin again will unpin the sidebar so it will retract when not in use.

Statistical Tests

This Statistical Tests option is available only when in the multivariate viewing mode, by selecting the Statistical Tests checkbox within the Visualization Options Sidebar. It provides additional detail describing the statistical relationships between the data underlying the chart in view.

Video Tutorial

Related Article

• Analysis Tab