Familiar. Collaborative. Managed.
In our next guest post, MVP Jen Stirrup provides an in-depth explanation of one of her favourite features of Power View. Read on and see how trellis charts can be used to help you get more insight from your data. You can also find Jen on Twitter @jenstirrup and follow her blog at http://www.jenstirrup.com.
I’ve got my 3D glasses. Now, where’s the 3D effects in Power View?
PowerView allows you to visualise data in different ways, such as the use of line charts and bar charts. One thing Power View does well is that it facilitates the presentation of the data easily and clearly. Take, for example, the situation where you’d like to display a lot of data on one graph; it can quickly become a mess of lines and colours on the page. This ‘mess’ is known as ‘chartjunk’, which was a term coined by Edward Tufte; essentially, it refers to noise or distractions, which serve to obscure the ‘message’ of the data.
One thing you'll notice about Power View is that it's doesn't allow you to create 3d charts. In my opinion, this is an extremely brave thing for the designers to do. I often see 3D charts, which can be perceived to make the visualisation more beautiful. On the other hand, 3D charts can make it more difficult to interpret graphs since they distort the visual relationships between the elements of the graph, for example, by foreshortening some areas whilst pushing others out towards the viewer. The viewer ends up having to do more work in order to understand the chart, since it is easier to understand a simple line, than it is to interpret a line with depth.
If you don’t have 3D Charts, does this restrict you from creating useful and interesting graphs? Well, Power View has the answer. One way of getting around 3d data visualisation is to use a trellis charts. Also known as a lattice chart, the trellis allows the data consumer to compare data by simply placing small charts in a grid fashion for the category variable. Put together, the charts can be read together in a comic-strip manner, since each individual chart block will tell its own story about the data. Here is an example of a trellis chart here, which aims to show the Gross Domestic Product, the Gross Financial Liability, and the Deficit for OECD countries.
In the above charts, the dimension ‘Country’ is a category or conditional variable, since the data in each chart is dependent on it. This facilitates the business user to review and analyse the relationships between the variables.
Since the charts are very simply laid out, the advantage of the trellis chart is that it facilitates the easy comparison of the countries’ performance. So, for example, we can compare the performance of Greece, with the performance of Estonia, by comparing vertically. Further, it is also possible to compare Greece and Germany’s data by horizontally comparing their charts.
It is also possible to examine the metrics within the charts, too. For example, we can see that Chile does not show much distinction between the GDP, GFL and Deficit metrics. On the other hand, we can see that Greece does show an increasing difference between Gross Financial Liabilities and the Gross Domestic Product.
One key item to trellis charts is that the axes should be synchronised – both horizontal and vertical. As Stephen Few (2009) points out in his book, Now You See It, ‘comparison is the beating heart of analysis’. In order to analyse data, it is essential that the comparison of data is facilitated, and Power View automatically synchronises the axes for you. The synchronisation of axes takes more steps in Reporting Services, and it is nice to see that Power View allows the user to produce synchronized axes so easily.
To summarise, trellis charts are a very nice, simple way of displaying the relationships between category variables, such as Country. Trellis charts are produced simply, and can be filtered in order to display the required data. Trellis charts allow the data consumer to review data between and within the charts – thereby facilitating comparison and a richer understanding of the data without the use of 3D.
When will trellis charts become a part of Excel? Combined with Analysis Services this would be a Tableau killer - one of the most significant competitors in analytics tools.