Did you know… Lumira smarts?
Lumira constantly aims to provide you with additional insight into your data… and often without you having to do anything more than click a mouse button! Two examples of this are the Manipulation Tools panel and Influence Analysis.
In the Prepare Room click on any data cell. The Manipulation Tools panel will appear on the right hand side. This panel offers many ways to alter your data, but of interest is the list of members and their associated record count that is automatically displayed.
Just by itself, the member record count provides valuable information. The above example is using a sales database so we can instantly see that Susan Larroux has made almost double the number of sales compared to Peter Rickardo.
Imagine that a geographic data cell was selected. You now instantly see a bar graph of the number of sales per geography, and sorting the Records column immediately shows the regions with the least and most sales.
However, depending on the selected data cell, the member record count can quickly provide useful insight and also provide data validation. One example is unique data where there should be no more than 1 record for each member. By choosing a descending sort for Records you can instantly confirm if this is true…
…and all of this information is derived from Lumira during data preparation, without adding any visualizations.
When exploring a new dataset, or if the dataset has a large number of dimensions or measures, it can be difficult to see how different dimensions and measures interact, i.e. how they influence each other. Lumira contains a very useful feature called Influence Analysis which can help in two ways.
1) In the Visualize room click on a measure, then click on the light bulb (Related Visualizations) icon. Some explanatory text will appear. At this point you can optionally select a dimension, but for initial exploration don’t click anything else… just click the “Run Analysis” button. The result is a list of visualizations in order of influence of the dimension, with the first visualization being an overview of all influencers.
In the above example you can see from the bar chart that the dimension with the most influence on the Commission Bonus measure is the “Missed Quota Count” dimension, followed closely by the “Hit Quota Count” dimension. We can also see that the “Month” dimension has limited influence on the Commission Bonus.
2) In the Visualize room click on the light bulb (Related Visualizations) icon. Lumira will automatically determine if there are related visualizations compared to the current visualization. These visualizations do not need to already exist… Lumira will dynamically create them if necessary.
Tip: Before viewing a related visualization, duplicate the current visualization by choosing Duplicate from the Settings menu. You can then keep the original visualization and the related visualization.
NOTE: If you do not see the light bulb icon, open File > Preferences and in the Datasets section make sure the option “Would you like to display Related Visualizations and allow influence analysis?” is checked, then restart Lumira.
Is this feature disabled in 1.31
This feature works even in v1.31. Just follow the same steps.
This feature is not available for HANA online connection
HANA online and Big Data connections are special scenarios with different approach strategies towards data. These two cases and the visualize room in BI platform never had this feature in the previous versions also.
Almost forgot about this feature! Maybe it should be listed more prominent when you start building your first visualizations after preparing your dataset. Thanks for the blog!
With kind regards,
I completely agree with you, as it would aid in building the rite visualizations. That is the reason I have included it in my series for beginners, Smarter Data Discovery using Lumira