IBM Developer Advocacy

Discover patterns, relationships, and predictive insights

Change column headings

When analyzing data in a Discovery, you may want to change the name of one of the column headings to something more meaningful. The following video will show:

  • how to change column headings in the data tray in a Discovery
  • the effect of changing column headings in the:
    • visualization
    • question section
    • data tray

Sort data

You can sort data in a discovery in ascending or descending order. You can sort the data in either the x-axis or y-axis, depending on what type of data is in an axis. The following video will show:

  • how to sort data items in a visualization numerically or alphabetically, in either ascending or descending order
  • the effect of sorting data items in a visualization numerically in ascending order

Leverage natural language in your analysis

In Watson Analytics you can pose natural language questions to help begin your analysis. Use the language and keywords of your business to build questions that explore and visualize your data. Use column headings and data values from your data sets, along with keywords to formulate your question. If you need help, use the question coach that provides suggestions for writing your question. The following video will show how to:

  • enter natural language questions on the Data tab
  • enter natural language questions on the Discover tab
  • use a starting point Discovery to perform analysis based on the question that has been entered

Navigate the spiral in a predictive analysis

A spiral visualization shows you the key drivers, or predictors, for a given target. The closer the driver is to the center, the stronger that driver is. Watson Analytics uses sophisticated algorithms to deliver highly interpretable insights that are based on complex modeling. You don’t have to know which statistical tests to run on your data. Watson Analytics picks the right tests for the data. You can add and explore a visualization for each key driver that will give you detailed information about what drives each behavior and outcome. The following video will show how to:

  • create a Discovery using a data set and examine recommended starting points
  • ask a question of your data and examine the new recommended starting points, or using your own visualization
  • choose the recommended Spiral visualization
  • examine predictive factors impacting the target:
    • in the spiral
    • in the table to the right of the spiral

Edit targets in a predictive analysis

After using a Spiral visualization and the default target to perform initial analysis, the target can be edited. The following video will show:

  • how to replace the target at the center of a Spiral visualization with a different target
  • different methods for replacing the target
  • how to replace the target using drag and drop from the data tray
  • how to change the aggregation type of the target

How to view decision rules and the decision tree in a predictive analysis

Decision rules are a set of statistically generated profiles, with each profile showing you a group of factors that are used that to classify records. This classification helps identify which combination of factors are probable to result in a specific outcome for the target field.
Decision rules can be interpreted using the Decision tree. The decision tree shows you patterns of characteristics that lead to a certain outcome, which again, you can think of as profiles. Each branch in the tree is a unique pattern that leads to the likelihood of an outcome occurring based on what has occurred in the past. Each level of the tree has a higher predictive strength than the subsequent levels and branches. The following video will show how to:

  • view and interpret decision rules
  • view decision rules in a decision tree

blog comments powered by Disqus