After a few days of learning and experimenting, I finally reached my goal to create a submission for the SAP Data Genius challenge. Armed with a LOT of demographics data and a some interesting ideas for analysis, I fired up SAP Lumira 1.28 and spent two days immersed in Lumira.
Along the way, I found a few black eyes, nice features, and tricks that I wanted to share as you start building storyboards.
What I learned after 2 days of Lumira Immersion
I think the biggest winners for Lumira are those users who spend hours meddling around in Excel pivot tables and copying/pasting visuals from Excel into PPT. For SAP BusinessObjects customers, these are the same users who export from Web Intelligence to Excel, before performing real analysis for business stakeholders. From my perspective, the role of a Lumira storyboards is a medium to highlight findings from data exploration tasks. Storyboards should not be mistaken for dashboards even though visually they are similar at a glance; functionally they are very different.
Tips to for your SAP Lumira Storyboard
Tip 1: If you use Excel, include extra columns with placeholder titles
At any point you can refresh your Excel document from the main menu: Data>Refresh Document.
However, I did find that new columns added to an Excel document are ignored by Lumira upon re-import. You could add a new file and merge / join the new data, but I found it easier to include 4-5 dummy columns. The reason I had to add new columns to my Excel were calculations that Lumira lacked like “median” and “rank.”
Tip 2: Start your storyboard with a title slide
With 1.28, you can reorder visualizations but you can not re-order storyboard pages yet. Make sure you have a title page placeholder because you can’t add one later. The good news is adding a new story is an easy process to re-create and arrange your pages. Hopefully upcoming versions of Lumira will grant more control to re-order pages.
Tip 3: Data visualization has to occur in 10K record batches
Originally, I wanted my analysis to be the most desireable zip codes to live in the United States. I had to slim my analysis to California because I found that 40K+ zip codes in the US wasn’t a good starting point with trellis charts. While it is difficult to visualize 10K+ records in a single chart, trellis charts can easily put you over the 10K mark, so be mindful of using filters effectively.
Tip 4: Use input controls for dimensions for basic dimension filtering
Input control provides interactivity so end users can select, search, and multi-select categories (dimensions).
Tip 5: Pay attention to design and composition
I offer the same advice I have offered for years to data artisans and business intelligence practitioners.
“Data visualization and analysis is not your medium to exercise your creativity”
- Avoid extra graphics and photos: For the purpose of a contest like Data Genius, you can expect more flash and sizzle with graphics for differentiation, but for day to day use, you want to
- Use text labels and titles: The chart titles do not have text wrap so if your titles are long, use text.
As we see updates and upgrades to Lumira some of these tips will get rolled into the software itself. If you have found a way around some of the issues I found or have more tips, feel free to share below.