Following on from my last blog Hidden Gems inside Lumira – Export Data in a Table here is another hidden gem when using Lumira.
So firstly, hands-up. The title isn’t quite what it seems.
I’m not going to discuss what Influencer Analysis in Lumira is. This is explained much better than I could do by Avni Savant last year in the blog Using Infinite Insights with SAP Lumira
What I will be discussing is how Influencer Analysis can be used a little bit differently to provide more impact – I just couldn’t think of a good title.
Whenever I have used Influencer Analysis in Lumira, I have used it on the complete dataset that I have been analysing. My data includes some measures and some dimensions, which are irrelevant for Influencer Analysis; i.e. long/latt reference information for cities and measures that I don’t want to include. Because I cannot filter the data before I do my Influencer Analysis my results won’t be completely correct and accurate.
Also, I use calculations in my analysis which unfortunately I cannot select and don’t work with Influencer Analysis (as per Lumira 1.23).
However, there is an easy way around this using a table and the “Create New Dataset” function.
First, I create a table which includes my measures and dimensions that I want to perform my Influencer Analysis on (including any filters) – see screenshot below.
Then using the “Create New Dataset” option found in the cog icon for the table, (see circled option in the screenshot above) Lumira creates a separate new dataset based upon this selection.
I can then go to this new dataset using the dataset selector – see screenshot below.
Lumira has created a fresh new data set with the measures, dimensions and filters which I can now use to perform my Influencer Analysis on – see screenshot below.
I now go to the lightbulb icon and select my measure to run Influencer Analysis on – see screenshot below.
I have now performed a much more accurate Influencer Analysis on my data using a calculation in a few seconds using a few simple mouse clicks. Something that I couldn’t do in my original dataset.
I hope that this hidden gem helps you in your Lumira analysis.