The Simplest of Tricks to Further Enable Qlikview Interactive Infographics. (*Still requires common sense)

Infographics; the thorniest subject in the world of data viz: are they art, are they telling the truth, do they have a use, do they engage the viewer, do they serve an agenda over the truth, or are they just for trendy media types etc. Reading many of the related blogs and articles out there it would seem that people fall into 2 camps; those who love Infographics (David McCandless) and those who hate them (Stephen Few) but for me the right place to be is to have a foot in both camps. That is to realize that Infographics all too often stray into the pseudo-practice of ‘Data Art’ and achieve nothing despite shouting loudly, but conversely when used correctly a little bit of ‘graphic’ can ensure users really engage with the information presented. So with that in mind I’ve been thinking recently how to use some of the better elements of Infographics in the Qlikview world as well as using Qlikview as a tool to create and enhance Infographics.

Before I get into too much detail I’ll clarify what I hold the difference between ‘Data Visualization’ (what Qlikview usually does) and ‘Infographic’ to be. A relatively simple explanation sums it up nicely; a Data Visualisation (Scatter, Bar, Pie Chart etc) can be used against many disparate datasets whilst on the whole an Infographic is tailored to a particular set of data and can’t be reused. Personally I feel a natural draw to Infographics, I’ve designed websites, logos and objects in the past so the design led nature of them appeals but the level headed lover of data in me pushes back; so many Infographics are nigh on meaningless and would be better off not even created, but I think there’s a happy medium where both sides can benefit so let’s try and find it.

The first thing I think when I see most Infographics no matter how good they may be is; ‘great…but it’s static, you may as well have drawn it by hand’. If only there were a software product we all know and love that was really good at intuitively and simply interacting with data? This is my ultimate aim; to create simple, intuitive, interactive Infographics that engage the viewer on a relatively narrow subject like a traditional static Infographic but then create further impact by allowing the ‘viewer’ to dive into the data in the traditional Qlikview way whilst always remaining true to solid data visualization design rules.

Firstly there have already been some steps along this road, the recent ‘My Life in Data’ app by Michael Antony at Qliktech used some layered objects to good effect, you can read my comments on it here: https://qvdesign.wordpress.com/2012/03/07/beautiful-design-chart-innovations-from-qliktech-and-the-elephant-in-the-dashboard/ (.qvw also available for download)

 So what’s this simple trick that makes Infographics possible in Qlikview? It’s the transparent .png image (along with a good dose of common sense of when and how to use it). As I’m sure you all know images can be imported into Qlikview via the Text Object, in most cases this might be a logo or map background, all well and good but it’s not much use here, what is of use is the feature that can be applied to .png image files in that they can be made to have a transparent background. It frees us a little from the constraints imposed by Qlikview – it is after all not designed to be a rival to Adobe Illustrator – and allows us to create more flexible representations of the data we’re looking at. I know that sounds like the most pointless tip ever but it really does open things up and not just for Infographics, below are a couple of examples of where I’ve used transparent .png files in the past:

A Subtle Use But an Effective One

The next one is admittedly crazy:

An Easily Created Annotation Layer

Simply put; by using a transparent .png you’re no longer beholden to match background color to what you’re layering over.

By way of an example of how we can do things with Infographics better in Qlikview I recently came across the graphic below from the Gaurdian’s excellent Data Blog about how people from a scientific background are under-represented in the UK House of Commons: http://www.guardian.co.uk/science/2012/may/11/kicking-down-doors-scientific-advice-governments

Now whilst I whole heartedly agree that it’s a bad thing that scientists are under represented I think there are misrepresentations that are even more important such as gender, age or ethnicity…or even just the way the population actually voted, so I’ve taken elements of the Gaurdian’s original and created my own Infographic in Qlikview using 2 transparent .png images.

A Slightly Interactive Qlikview Version

It won’t tale too long for you to work out how this has been achieved; it’s simply a single stacked horizontal bar chart overlaid with a transparent .png image where the shape of the House of Commons has been removed. It’s all a matter of lining everything up and setting the layer order correctly. The .qvw is available for download below so I won’t go into the details as you can see for your self but I will explain how I create my .png files so you can easily create your own.

Creating the Transparent .png:

There are numerous ways this can be achieved but I’ll outline the one I’ve used here as it’s free and available via your browser. For my general image / drawing needs I tend to use Sumo Paint a browser based Photoshop style application that’s easy to use and delivers good results – www.SumoPaint.com/app if you know how to use Photoshop you’ll be right at home. (Great for on site consultants who need a bit of quick image editing but can’t install anything on a client’s machine – it’s got me out of trouble more than once)

All the .png’s I create start with the required Transparent background; when selecting File > New Image ensure than you select the Background to be ‘Transparent’ instead of White. In the case of my Houses of Parliament image I’ve taken the Guardian original, removed the text via copy + pasting blocks of color from the original over the areas that aren’t required- ie to remove the colored graduations and have then flattened the image. That leaves us with an all grey Houses of Parliament graphic, we can then use the Magic Wand tool to highlight the various sections of grey before Cutting them from the image to reveal the transparent background behind. Once complete we have our overlay ready to go over the top of our single bar chart to complete the Infographic.

Of course there are a myriad of alternative methods that could be used to achieve an endless list of Infographics it’s really down to your imagination, below is a similar example I’m working on ready for the US Election:

Now there will be a number of you out there shouting: ‘but how’s that useful in a business dashboard!?’ and the answer may simply be; it isn’t. That said that doesn’t mean we shouldn’t be creating Infographics in Qlikview, not every Qlikview dataset has to revolve around stodgy reports and sales figures, so set yourself free and visualize an interesting data set in an interesting way just for the love of data!

There will also be a good number of you out there thinking: ‘I can’t be bothered with that it takes ages to fiddle with the .png, line everything up, polish the layout etc etc all for one visualisation’ and in a way I agree but you’d be missing a fundamental point. Creating Infographics in Qlikview does take ages…when compared to creating a standard out of the box Qlikview dashboard, but just think how long it would take to create my Houses of Parliament Infographic in say Silverlight, HTML-5 or Flash (all of which would be the natural choice); it took me 1.5hrs with Sumo Paint, Excel and Qlikview start to finish.

This is something I’m going to continue working on so I’ll keep posting examples of what I come up with; the next step being to build one with true data interaction so watch this space.

Finally a plea to Qliktech for some Qlikview.next functionality; please create something similar to Tableau Public whereby I could create a Qlikview based Infographic, publish it to a cloud based Qliktech owned Qlikview Server for free and then embed it in a website or at the very least link to it. I think this would really expand the awareness of Qlikview and what it’s capable of by getting it’s functionality out there on websites and blogs. The Qlikview way of doing data interaction should be the natural, standard, expected way of doing data interaction and to get there it needs to break free from the BI / MI world and get into the main stream and the accessible thought provoking Infographic is the perfect way to do this. Of course there would have to be limits to the dataset size etc to avoid cannibalizing Qliktech’s license revenue but I don’t see any problems with it and seemingly nor do Tableau.

The .qvw for the MPs Vs Everybody Else app can be downloaded here: https://docs.google.com/open?id=0BxloTMUod74tRkxfSnFfM2ZIejQ

As always; all the best,

Matt

Advertisements
Comments
2 Responses to “The Simplest of Tricks to Further Enable Qlikview Interactive Infographics. (*Still requires common sense)”
  1. More often than not, infographics are either needlessly obtuse (#1), or directly misleading (#2).

    Case in point #1:

    In the “vs Everyone Else” diagram, comparing “How it is” with “How it should be” is the whole point of the graphic. Unfortunately it puts the two horizontal stacked bar diagrams side by side, making it very hard to compare the sizes of the blocks. Simply putting one diagram above the other would significantly improve comparisons, without loosing any visual impact.

    Case in point #2:

    The “650 MPs” diagram uses an artsy stacked bar chart. The misleading problem is that (geekily counting pixel widths:-) the stacked bar chart simply adds 1 + 13 + 60 seats to a total of 74 seats. Unless I’m misreading something, there is nothing in the real world corresponding to these 74 seats, so the graph overstates the proportional number of science workers. Rather, the 1 and the 13 seats would be either separate from or part of the 60 seats case – a non-stacked bar chart would likely be a good choice.

    My advice to myself when prettifying data is to always start with a solid data visualization that portrays the data clearly, and only then add the artsy icing on the cake (making sure it doesn’t misrepresent the original visualization.)

    Cheers!
    Kristian

    • qvdesign says:

      Kristian,

      I don’t entirely disagree about Point #1. I did consider placing the 2 charts vertically to aid granular comparison as well as possibly inverting the 2nd chart to ensure the constant bottom edges were next to each other to potentially make comparison easier.

      My reasons for not doing so were:

      1.
      This isn’t a data visualization, the aim isn’t to compare say 12 seats for Labour in one chart with 13 in the other, it’s to show very wide disparities between the 2 charts and as such I don’t feel that having them side by side really takes away from that. The aim is to try and convey a more general message (‘The Liberal Democrats are underrepresented’ not ‘The Liberal Democrats are under represented by 34 seats’) and make it as accessible & meaningful as possible Had it been about conveying exact numbers then the 2 charts wouldn’t have been side by side nor would they have been shaped like the House of Commons and there would more likely than not only have been a single chart containing both data sets to make accurate comparison as easy as possible.

      2.
      Secondly the spire of the chart would have meant had the 2 charts been aligned vertically there would have been a large gap between the majority of the 2 charts’ data areas making accurate comparison harder than it could be, this is further compounded by the fact that as the data is shown via a stacked bar chart the sections wouldn’t line up for comparison anyway. This is a drawback with the usage of the House of Commons shape as it impedes the vertical alignment and forces the use of a stacked bar however I feel that without that device the Infographic would loose much of it’s resonance and connection to the source data.

      3.
      When people compare something in the real world they tend to place them side by side (X-Rays being a prime example), our eyes are side by side after all, it’s simply more natural and intuitive to place them that way when trying to get an impression and not needing to take exact readings.

      I do agree that if an Infographic can both provide the visually resonant connection to the data source, be easy to access the general message and also show the detailed data as good as a pure data visualization then that’s the best solution. However, in cases like this where making the purest data visualization would loose the resonance and going too far with the visuals would obfuscate the underlying data it’s about finding a balance between the two.

      As for your second point; there’s no excuse for the Guardian Data Blog using that trick to exaggerate the disparity in the situation, it’s either a school-boy error (accidentally stacking the chart) or an intentional attempt to mislead – both of which aren’t acceptable. I guess this is one of the major pitfalls of Infographics; if it’s hard to read the data to it’s lowest level then it’s hard to verify that the ‘general impression’ it’s providing is actually accurate, in those cases you need to be able to trust the author.

      All the best,

      Matt

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: