In the following example we work through a real world example of a data visualization. We’ve chosen an example that involves Operations data – this is fairly non-domain specific so hopefully it can demonstrate some important points. The first, and most important point is that you have to define your audience.
We receive many questions about “what is the best chart for this situation” or “what colour should I use for emphasis”. These questions are usually attacking the problem from the wrong angle. The one question you need to ask before anything else is “who is this visualization going to be seen by and how?” Is it in a boardroom on a printed sheet or across a trading floor on a plasma screen. Are the consumers domain experts?
This example features data about an investment bank’s operations processing, the audience being the clients of the Operations department.
Initially the project started out as simply trying to record what operational problems were encountered on a daily basis across different product lines. A reporting system was built and various generic reports produced:
Unfortunately the reports either didn’t contain data at a granular enough level or it was difficult for the product managers to see where the issues were occurring and what the trends were. In reality the report showed what the major problems had been – unfortunately this was already known, as when something major goes wrong you remember getting shouted at!
What was requested
The client wanted a report that showed where the problems were occurring across business lines (rather than operational units) and how they were doing historically in a single page that could be included in a weekly MIS pack (they currently had four pages per product line (8) so a total of 32 pages. As a first pass they simply wanted an Excel worksheet they could update manually:
We felt this solution lacked clarity and it was very difficult to spot trends across products.
What we proposed
We designed a solution using MicroCharts to allow small multiples of charts to show a variety of views:
This solution allowed the user to view the data simply as a cumulative set of data by Product (top line) or by Root Cause (vertically) and then look deeper into historical trends in the centre of the chart. For example, its fairly easy to see spikes in the Root Cause data historically and see that the overall trend has improved over time. By ranking the Products and Root Causes you immediately give some sense of scale to the data. For example you can see that there are many more Application failures than any other type of problem, but the majority of root causes are otherwise fairly evenly distributed.
One other point worth noting was that the original colour scheme was much more muted, but the client got very upset that it looked like a competitor’s corporate colour and wanted it to be “louder”.
What was the user reaction…
Ecstatic, 1 page replaced 34 and they could see at a glance how the entire (large) organisation was working but also quickly find out detail for a particular area and identify trends.
3 Replies to “Data Visualization – a real world example”
It’s traditional to put the Total row on the bottom, but I think here it might have been better to have it at the top near the colored bars. Might there have been a little extra value in a Total sparkline column on the right, next to the “share by product” stacked bars?
Very nice example. I find your solution much more “workable” because you don’t have such a strong bias on the FX & MM column.
Thanks for the comments guys.
We did play around quite a bit with the layout, and one version had the total sparkline on the right also, but in practice it seemed to lose the framing effect which the bars give.
The main reason we went with this version though was down to the business, in that it’s the one they chose from 3 slightly varying alternatives we offered!