Qlik Summer School Webinar Series - Visual Analitics Design Best Practices

The Visual Analytics Design Best Practices webinar was presented by Patrick Lundblad, Senior Product Manager, and Adam Booth, Product Designer. This session was really interesting not only because it was full of tips about data visualization design, but also because the very interesting facts about data visualization history.

So, let's start with some of the historical facts presented in this webinar. Data analysis techniques have been applied in a lot of fields (such as economics and medicine) in the past few centuries. It was interesting to know some of the names of the pioneers in this area like Charles Joseph Minard, John Snow, Florence Nightingale, Joseph Priestley, and William Playfair who according to historians created the first bar, line and pie charts (indeed I was so fascinated about them that I've created a separate entry).  Also, it was interesting to know how the London underground map has evolved to provide a more useful version that is easier to understand by users.

At the beginning the map shown the names of the stations and their locations, but it was difficult to understand due to the size of letters, the position of the names, and even in some areas near to the center the name of the stations were omitted. Later, around 1931 the new version of the tube map, remove the geography of the map and focused on show the position of the stations and how each of them were related. This new version was cleaner and easier to read and included more information (and since then the map has been evolving for good). The lesson here, check what has worked or hasn't worked over the years, and use this information to provide better and more useful visualizations.

Now, let's move on to best practices. One of the recommendations during the webinar was to think on three different things when creating data visualizations: Data attributes, visual encoding and usage.
  • Data attributes: For instance, if the data is discrete (size, position, sex, sports) or continuous (costs, age, temperature, time) *
  • Visual encoding: How to represent the information using colors, shapes, size, or position.**
  • Usage: What's the purpose, what are you trying to show or demonstrate? Here is a visual help to select the right chart***


Also, additional considerations when creating charts:
  • Use the color to make things obvious, for instance red for numbers not meeting the target and green for the good ones, or you can use symbols
  • Beware of color blindness, some people don't see certain colors and distinguish cues
  • For more important values use different colors
  • Choose the right dimension and right granularity
For more information about best practices, I recommend to visit Patrick Lundblad's blog, I have also added in the footnote the links to his posts about the three pillars of mapping data to visualization (Data attributes, visual encoding, and usage).

* Patrick has written a very detailed post about data attributes here. Also during his presentation, he suggested the Community Design Blog - Scales of Measurements
** Refer to the article Second Pillar of Mapping Data to Visualizations: Visual Encoding for more information.





No comments:

Post a Comment