Cloud analytics has been grabbing me by the throat in a very positive way. Week after week, I’m discovering the latest capabilities, and week after week I’m surprised by how far cloud analytics has already come. I already summarized some of the highlights in an earlier blog, but there are so many I would like to write this Part Two. And again, it’s about a number of best practices and tutorials I believe are important to understand.

In today’s “roadtrip” with SAP BusinessObjects Cloud, I cover the best practices of variance reporting, automated storyboard creation, cross-chart filtering, decision trees, and predictive forecasting.

Variance Reporting

You probably already know that one of the strong points of SAP BusinessObjects Cloud is that is supports the Closed Loop Portfolio of analytics. In one application we can combine business intelligence (monitoring) with business planning (budgeting) and predictive data. Even more, we can simulate how an adjustment to our planning would affect our monitoring and more. To do so, we need to populate our SAP BusinessObjects Cloud model with not only “actual” metrics, but also with our budget data and/or initial forecast. If we have done that, we can create variance reporting (with Hichert – IBCS support) comparing actuals, budget, planning, and forecast figures on the fly.

 The “trick” here is in correctly loading your data into the SAP BusinessObjects Cloud model, which I explain in detail at this page.

Automated Storyboard Creation

 Imagine you get access to a new data source and quickly want to know what kind of information it stores. Well then, the automated story-generator within SAP BusinessObjects Cloud is for you! In three (!) clicks a dashboard automatically generates for you that pitches an overview of your data. A few clicks more to customize it to your needs, and it delivers you a professional dashboard. Don’t know how they do it, but this one rocks! (If you don’t believe me, look watch this.)

 Cross-Chart Filtering

Everybody loves the cross-chart filtering facility we know from SAP BusinessObjects Design Studio. Did you know it’s also available in SAP BusinessObjects Cloud? (This video shows how to do it.) The connection between various graphs is driven by the Linked Analysis option. One can choose which of the graphs in your story are affected or even the whole story (including subpages). One step further is the Data Point option. If you apply this option, the applicable attribute-value(s) filters all of the dependent graphs immediately. This is very useful and for me one of the key visualization options that makes me love SAP BusinessObjects Cloud.

Cross-Chart Filtering

Decision Trees and Predictive Forecasting

The decision tree applies a predictive model on your data that shows the correlation of your data’s attributes towards the measures. In other words, imagine you want to understand which of your attributes is the strongest influencer on your revenue. Is it the region where your products are sold? Is it the weather, or is it the product category? In the video on this page, I use repair cost as a metric to finally discover if an unexpected equipment category is the main influencer. The outcome of the model is displayed in various ways, including correlation information, various graphs, and charts that can be re-used in your dashboard.

The decision tree not only can be applied to metrics, it can also be applied to dimensions. In that same video, we use it to create a base and comparison group. Then we use our algorithm to find out that a specific manufacturer of spare parts is our main influencer for the majority of our repair costs. This leads to informed decision making, since we might want to replace this manufacturer and buy our spare parts somewhere else.

​Predictive forecasting is possible using another predictive algorithm that is embedded in SAP BusinessObjects Cloud. Planning-based models are the source of the predictive forecasting that further needs a time dimension (monthly as lowest grain) and some historic data to run its triple exponential predictions. The outcome can be applied in a dedicated version of your data and drive your rolling forecast. If that’s not innovation!

For More Information

 

This blog first appeared on Iver van de Zand’s blog and has been republished with permission.