On a recent visit to a local watering hole, I was pleasantly surprised and delighted to find that this microbrewery establishment was quite data driven. Before I go into further detail, you may have had similar experiences when you order a drink after work on a Friday and the server comes back empty-handed with the look of gloom on their face. They have run out of the beer you were craving and the whole night goes to shambles…well maybe that is a little too dramatic.
Wouldn’t it be awesome to avoid such situations where you were able to see how much beer they have available so that you order a beer that is on tap!? Well this establishment, on Commercial Drive in Vancouver, BC called St Augustine’s Pub, has done just that. As you walk into the establishment, you are greeted by a 40″ TV over the bar which displays their beer selection visualization and how much is available of each brew. I applaud the effort in having something like this available but I do have to comment on the execution of the visualization and what I would consider as improvements on the design to make it more user-friendly.
Since I am such a data viz geek, I took some photos of the beer dashboard to show you what they created which will make it easier for me to comment on the changes I would suggest. In doing so I will also bring up some key concepts in data visualizations which you should keep in mind no matter if you are creating a beer dashboard or a dashboard for key indicators for a Fortune 500 company.
As the title alludes to, the key graph used to display the amount of beer left is the very controversial pie chart. For those new to data visualizations just google pie charts and I am sure you will find some of the articles very against pie charts and message boards full of everyone’s opinions on the matter. You may ask where do I stand on the matter of pie charts… well visit my vizism post on pie charts.
The original beer dashboard at St. Augustine's Pub
Back to our beer dashboard, as you can see in our first screenshot there is a lot of beer to choose from! I love trying new brews, so it took me a few moments to orient myself to the beer dashboard. As a customer you can figure out a few things by looking at the visuals right off the bat:
- Color of the beer in each glass allows you to pinpoint a lager from a stout quite easily
- The names of the beers and breweries are clearly displayed in big font on top
- Any beers unavailable are clearly overlaid with the word empty
Close up of the skewed pie chart
That is great and makes it easy to pick beers but what could be improved is the visual for the remaining beer. They have created a three-dimensional pie chart using the rim of the glass as the circle or “pie”. If you look at the St Augustine’s Lager beer (which is so refreshing in my opinion) below you can see the problem with using pie charts in this form. It looks like there is 50% beer left but actually there is 35% left! For us mere humans we have a difficult time getting exact numbers from a two-dimensional pie chart and now with a skewed three-dimensional one you might as well just display a full glass and rely on the numbers that are located on the bottom right of each tile to get the job done.
What can we do to improve on some of the shortcomings? Well let’s break it down.
I would ditch the pie chart and change the glass so it shows the entire glass on the bottom right of the tile. I moved the brewery and beer names over to accommodate the new glass visual. At a glance you can now see how much beer is left without having to do geometry. You also have more of the color of the beer showing in this design then the unsightly foamy head so you can figure out what kind of beer it is by the colour alone.
First crack at the new beer dashboard widget
Always have a mindset of figuring out what is the story you would like to tell with a dashboard and avoid creating a series of novels
I think these are pretty good improvements but what would take it to the next level would
be changing it from merely showing how much quantity is left to showing how many pints are left! Now you won’t have that situation of trying to order a beer which has run dry! So I have replaced the % remaining with # of pints left. In doing this we have taken the guess work of what does 35% amount to in consumable pints? This may not be an issue when you have 35% left but for those that have 10% or less it may mean that there are only a few pints left! We have taken a step forward in the thinking process to include pints remaining, which you can do with any other dashboard by looking for areas where you can remove a barrier to the data by simplifying what you are displaying.
Second crack at the beer dashboard widget showing pints remaining!
My version compared to the pub’s version is slightly different but improves on it so that you can get more information without the need to interpret it while you are enjoying the company of friends and pints. Now that is a visual I can raise a cold beer too!
Make sure to visit St. Augustine’s website where you can navigate and browse their live beer menu which is similar to the dashboard at their establishment.
What do you think of having visuals like these in pubs? What other types of establishments do you think would benefit from a dashboard or data visualization? Post your ideas in the comment section below.
Nice write up of a fine establishment! A few nits to pick:
Regarding your redesign:
a) By polluting the right side of your widget with the full pint glass, you’ve cut the text-space by nearly 40%. Works fine for the lager, but what about Rogue’s Imperial Younger’s Special Bitter? And even for ‘Howe Sound Brewing’ you’re running the risk of stomping the pint, or awkwardly stacking text. (Reading left to right lends well to the horizontal…)
b) I think the full pint view is a bit…overt. The skewed pie chart has a touch of subtlety you only really get to indulge in when in the data is less important — and honestly, the actual amount remaining IS less important than you let on. At the end of the day, you shouldn’t care how much is remaining, just that there IS some remaining (though, see point (3) below for the twist I’d put on it) – and the ‘full vs empty’ effect of the pie chart communicates this fairly well. (My eye immediately picks up the white space difference between, say, one of the stouts and the Loki)
c) The pint measure, while far more practical than a percentage, has a couple holes. St. Augustine’s offers a ‘tasting paddle’ which lets you sample 4oz of 4 different beers, as well as offering sleeves. So you’d either have to go fractional (which no one likes) or round down. Further, since pouring a pint is more art than science, you run of risk of misinformation. The percentage system gives you a built-in fuzziness that covers for spillage, miscalibration and the odd sizings.
3) To the point of important information – the one piece of information that I would love to see is the ETD – estimated time of depletion. By tracking the rate at which a beer is being consumed, you could project the time at which the beer will run out. A beer at 10% isn’t in danger of not being there for your next round if no one else in the place is ordering it…but the beer at 30% could be gone if they’re pulling a pint a minute…
At the end of the day, we do agree on the important points: The real time display is awesome, the clean lines and color choices impart the important information quickly and beer is only made better by data about beer.
Cheers!
Hi Scott,
Thanks for your comments. A lot of valid points here so I will try to cover them all.
Your suggestion of Estimated Time of Beer Depletion is an awesome idea and I can imagine that being displayed as a sparkline showing the depletion rate for the last 2 or 4 hours. I plan to make a post on sparklines outlining what those are but they are essentially small concise line graphs with very little labelling. I think of it as a more detailed trending arrow that you usually see being used on scorecards and is standard feature in Microsoft Excel 2007 and onwards.
Thanks again for your insightful comment
Cheers,
Dan