Brand attribute breakdown
Cole Nussbaumer Knaflic, Storytelling with data, is one of the most appreciated young voices in Data Visualization. You can read more about her work in this interview. Her blog, mentioned in almost any top of Data Visualization Blogs, will be worth your time if you want to improve your knowledge and skills in Data Visualization. Just have a look at it.
Her recent entry, novel vs. the boring old bar chart, is a nice story about a team seeking for a novel way of showing data. The goal of the visualization was to compare the performance of a company against its competitor across several criteria. Starting from a sketch, Cole proposed three designs one using DOTS, one using SLOPEGRAPHS and another one using BARS.
I need to mention that this subject is very dear to me, I wrote an article about multi criteria analysis 25 years ago in the process of getting a Phd in Mechanisms. Short time after that I totally switched my focus from the mechanical field to data analysis, so I regrettably ended my ongoing Phd. However, 10 years later I came across the very same subject when I developed and implemented Balanced Scorecard software in a manufacturing company in Belgium. To make the story short, an usual multi criteria analysis is based on the relevance/importance of the criteria and uses an overall score to summarize the performance. These elements were missing from her design, so I decided to add fictive relevance to the criteria and the overall score for my own design. I went even further to present a simulated scenario showing an what-if situation in which all issues were solved. The below design uses the circle size to encode the criteria relevance, adds an overall and an what-if ideal score, making use of the same horizontal scale. The area below target has been filled with light gray. Needless to say that except the fictive comments on the right, this design can be adjusted for any possible scenario without changes. This is one of the major prerequisites for a good design, but unfortunately very often "omitted" by data designers.
For a description of the context, I invite you to read the original post. Once you are familiar with it, you may see the changes I made to her DOT version, scale adjustment, clean up of value labels, added attribute relevance encoding, overall score and what-if conclusions. The overall score is just the weighted mean or median of attributes scores. If all attributes would be equally important, then all circles would have the same size. Overall score calculation would be simple mean or median.
PS. Nick Desbarats, also decided to make his own design using multiple bar chart. You can read about it at his Practical Reporting Blog.
Owner, Logbis
8yThank you, Nick. As you noticed, my slightly different angle was related to my past experience. Truth is that we interpret data by checking patterns, trends, outliers and so on to draw conclusions. One obvious conclusion of the original sketch was that overall our company is doing better than our competitor. My extended design emphasises just that. If Cole story was indeed about a team seeking for novelty, overall score and what-if perspective approach might be just the touch they were looking for.
Interesting post, Daniel (and thanks for the shout out). I'm also very interested in composite scores/indices and will be writing about these in the future. As with any visualization, your, my, and Cole's designs should all be evaluated in the context of what the author's purpose was in creating it in the first place. When I designed my solution, I based it on the stated purpose in Cole's post, which was to draw attention to the one attribute where the competitor was beating our company, and the one attribute where we were below target. I think that your solution would work well if the goal were to provide a general overview of how our company is performing relative to its targets and competitors. In that case, it would make sense, for example, for the focus to be on the absolute values and less on the differences between values, and to include a composite score. Love the Zvinca plot, by the way. Nice work.