Let there be light: Data visualization with SAP Lumira

GIS and Business Intelligence (BI) are buzzwords you hear together increasingly often (see also our articles on GISconnector, which we consider a low-cost, easy-entry BI solution). Inspired by this article from iX magazine I decided to have a look at the „self-service BI“ solution SAP Lumira. Lumira is an analysis and visualization tool and SAP offers a freely downloadable version. There is also a standard edition with more features that sets you back almost 1,000 $. The workflow in the application is simple: You import some data and prepare it, create visualizations and publish them.


Data Source for all visualizations: World Bank (© 2010 The World Bank Group)

In the free version you can use Excel Sheets, CSV data, copy data from the clipboard or connect to a SAP HANA One database. The full version also lets you access databases with SQL queries. For my trial I found an Excel table about global energy use provided by the World Bank (downloaded from www.visualizing.org). After loading the data you can prepare it for the visualization: Filter, calculate new values or join different datasets together.

You can also create a so-called geographic hierarchy to visualize the data on a map. A geographic hierarchy is a kind of geocoding: Geographic placenames are matched with an internal database to add location to your data records. The good thing is, that it clearly identifies records that could not be automatically matched. It then offers you the possibility to select possible matches from a list. Unfortunately, you can only choose from this non-extendable list. For some reason it did not suggest Slovakia as match for Slovak Republic which would have left changing the country name in the input data as the only remaining option. Luckily, I had also country codes in my dataset which worked much better and are obviously better practice, if you have them available.



Now comes the really cool part: Drag and Drop visualization. Just select one of the many available charts, drop your measures and dimensions on the X and Y axes, apply some filters on the data and your diagram is ready. This is really comfortable (and I can tell, having recently spent hours producing some rather simple graphs with R and ggplot).

Apart from classic graphs you can also create maps. These are nice for a first impression, offering zooming, panning and mouseover information. But overall the maps are pretty basic with almost no options to influence the display. The full version offers the possibility to use ArcGIS Online maps which brings a broader range of functionality.
For my trial I tested only simple visualizations. But Lumira also offers some fancier variants e.g. heatmaps, network diagrams or tagclouds.

After creating visualizations to make your point, you can aggregate them into a report or a „board“. The nice thing is that the charts remain interactive. I haven’t yet tried all possibilities but you could probably do similar things as with storymaps.

The next step is to make your visualizations available for others. Unfortunately there is (at least in the free version) no option to export the graphics as PDF or image files. This would be useful in order to be able to include the graphics in reports and presentations.

One possible solution to this is to upload your board to the Lumira Cloud. That is very neat and you can then provide access to individual users or the whole world. proved to be an unexpected hassle and you can only enjoy my screenshots for the moment.

A few words about the data

For this little test I was more interested in the tool than in the data itself. Some things look quite interesting though. Whereas the per-capita energy consumption for most countries has increased between 1970 and 2005, Luxembourg shows a massive decrease. My first hypothesis for the cause of this unexpected outcome is the demise of the heavy industry, but I have not yet found a confirmation for this. Perhaps you, dear reader, know more?