Skip to Content
Author's profile photo Flavia Moser

Guided Machine Discovery on SAP BusinessObjects Cloud for predictive analytics

What is Guided Machine Discovery?

Guided Machine Discovery (GMD) is a new set of capabilities in SAP BusinessObjects Cloud for predictive analytics that is powered by the SAP HCP predictive services. The guided analysis is designed for Business Users and features the power of Exploratory Analytics by leveraging SAP’s proprietary predictive technology. GMD allows business users to take advantage of predictive analytics without the need for any Data Science or Machine Learning expertise.  It enables business users to interact with insights, in form of intuitive charts and natural language, in order to make faster decisions and share newfound valuable insights with their organization. Let’s have a look at a concrete scenario.


Scenario: Guided Machine Discovery Automatically Identifies Key Influencers to Increase Sales

To demonstrate the user-friendly GMD feature in SAP BusinessObjects Cloud, this is a scenario where an organization hires a lot of contractors for their work. They would like to know what has a positive and negative impact on whether or not their contractors complete their assignments above or below their given budget. Through GMD, this organization will be able to discover the influencing factors and then can take action to ensure contractors finish within their budgets.


Setting Up:

  • Create a “New Story” and import your data
  • Create a “New Machine Discovery”



For this scenario, the “Percent_of_Budget_Consumed” column in the data shows how much of a contractor’s given budget was used for their project. We are looking to see what data is influencing the “Percent_of_Budget_Consumed” field. GMD’s interface allows us to easily exclude measures and dimensions from the analysis and we will be excluding “Budget_On_Track” from our analysis in this scenario.

  • Let’s select “Percent_of_Budget_Consumed” as the measure we are looking at
  • Uncheck the box next to “Budget_on_Track” to exclude it from our analysis
  • Change the Name of Records to “Contractors” as each record corresponds to a contractor in our data
  • Click Run
    This is Guided Machine Discovery running predictive technologies with the different algorithmic models in the background to help us find the one that fits our data best.




We can take a look at the Insight Quality at the top of our page to see what our data tells us. The engine that produces these insights is extremely robust and will only produce insights if it is accurate. Here is what we see for this scenario:


Result Overview

So what did we learn? Well, we know that the insight quality is rated 4/5 which is great. In addition, of the 19 possible columns we provided for analysis, 10 of those have come back as Key Influencers. We can also see the column we chose to exclude and filters applied if there are any. Now as business users, we can take a closer look at what these influencers are to see what further insights we can gain from them.




We can see the bar chart on the left displays our top influencers of the “Percent_of_budget_consumed” and by clicking into them, we can gain further insights to help us learn more about specific influencers’ impact. It turns out “Supervisors” have the highest influence on whether or not contractors finish on track with their budget. We can take a look at this interesting insight and see what we can learn from this. It looks like Udovicic’s contractors consume around 20% more than the proposed budget. This is really high and it is definitely something the organization should take action on!

Now we want to know if there is a correlation between the “Supervisor” and “Contractors’ Ratings”. Below the two bar charts we are looking at now, we can also visualize the analysis and compare different Key Influencers combinations. These metrics can show us interesting combination of fields and insights.

  • We already have “Supervisor_Name”  selected, now we can also click on “Rating” to look at how both interact with “Percent_Of_Budget_Consumed”.




When comparing the supervisor to the rating, through this heat map we can see that Udovicic’s contractors tend to finish above average budget consumed despite the type of rating the contractors had. We can now use this information to keep an eye on Udovicic and look into why their contractors are finishing way over budget.

Lastly, after gaining and gathering all these new insights through our analysis, we would want to share this information with our colleagues. We can simply select any of the charts we looked at and pin them to any of pages for sharing. 

  • Click on any chart we have created and want to share
  • Click “Copy to Page”
  • Select the desired page to pin the chart for sharing


Assigned Tags

      You must be Logged on to comment or reply to a post.
      Author's profile photo Senthilkumar VP
      Senthilkumar VP

      Great ! Good One.. collaborative.

      Thanks for Sharing.

      Author's profile photo Former Member
      Former Member

      The simplicity is impressive.

      What is the minimum amount of records necessary for the predictive algorithmic models to work?