Beer: Finding Your Favorite That You Didn't Know About!

Data analysis can be used for many things . . . how about finding other beers you might like, so you don't keep drinking the same old brand every time? Hang on tight -- I think we're about to make a beer run!

I recently read an interesting article on the Flowingdata website, where they graphically charted 100 beer styles. For each style, they drew a rectangle with the width representing the amount of alcohol by volume, and the height representing the bitterness (hoppiness). They colored the rectangle to try to represent the average color of the beer, and grouped the graphs by family. As you mouse over each of their graphs, it gives you a description of that style, and lists several different brands of beer from that style. Here's a screen-capture of the graphs for the family of beers in the Pilsner style, for example:

I found their graphs very interesting, but I also noticed a few things I would have done differently, using SAS graphs. Let me walk you through my changes and enhancements, and see if you like them!

One thing that I found baffling was that they showed an overlay of all the beer style rectangles at the top of the Flowingdata article, but it was purely for artistic purposes. It had no axes or grid lines, and there was no way to tell which rectangle represented which style.

In my version, I used overlay graphs for their analytic power (rather than artistic power). I created an overlay graph for each style family, so you could see how consistent (or inconsistent) the beers from that family are. For example, here's my overlay graph, followed by the individual style graphs, for the Pilsner style family:


In the Flowingdata article, they omit the text and numeric labels along the axes of the individual style graphs, and just show the labels on a single graph at the top of the article. I found that I had to keep scrolling back up to the top to see what the axis values were, and then scrolling back down to the style graph I had been looking at (and hoping that I had correctly remembered the values). By comparison, in my version I fully labeled every graph - this makes them a little more cluttered, but a lot more usable.

The graphs were very small in the Flowingdata article, and therefore the data rectangles were sometimes just a visual 'speck' with more of the black border color than internal yellow/amber beer color. I made my graphs about twice as big, to allow you to see the data better. And on the topic of color -- I decided to make all my polygons the same color, to make them easier to compare. (I'm not sure that an average color for a particular beer style is very valuable to graph, and I wonder if the colors in the original graphs are actually representative of the beer colors?) Also, the lighter and darker rectangles in the Flowingdata graphs could distort the visual perception of their sizes.

In their article, there was no way to navigate through the style families. You had to scroll up/down, and read all the family names, to find the family you were interested in (and the difficulty was compounded, because the names were not in alphabetical order). In my version, I create a list of all the style families, and let you click the style name to jump directly to those graphs.


When you hover your mouse over my graphs, you see the description of that style and list of several different brands of beer that are that style (similar to the Flowingdata graphs) . . . but you can also click my graphs to launch a Google search for that beer style. The Google search returns some really nice information, and also pictures of the beer (I think the pictures provide much more accurate colors than the colors used in the Flowingdata polygons, if you want to really know what the beer looks like). And for a finishing touch, I add a footnote at the bottom of my graph, giving credit to the data source, and a link to the actual spreadsheet containing the data.

And now, with all this data, how might you use it to find new/different beers, similar to the ones you like? I invite you to tell me in the comments section!

And what would my blog posts be without some randomly-related pictures from of my friends?!? This time, pictures of beer! . . . or should that be 'pitchers' of beer!?! LOL (Thanks Beth, Paul, and Jason!)

(Editor's Note: to see the original friend beer images as separate images, please visit the original SAS Learning Post blog.)

This content was reposted from the SAS Learning Post. Go there to view the original.

Robert Allison, The Graph Guy!, SAS

Robert Allison has worked at SAS for more than 20 years and is perhaps the foremost expert in creating custom graphs using SAS/GRAPH. His educational background is in computer science, and he holds a BS, MS, and PhD from North Carolina State University. He is the author of several conference papers, has won a few graphic competitions, and has written a book calledSAS/GRAPH: Beyond the Basics.

When Are the Fall Leaves at Their Peak?

Want to travel the country and see the peak fall leaves wherever you go? We've got a data visualization for you.

Graphing Mistakes to Avoid Like the Plague!

Plague outbreaks may hearken back to another era, but there are still instances of this disease today. Here's a data visualization that tells you more.

Re: Beer thirst
  • 1/12/2017 9:12:38 AM


Ariella writes "I'm still fairly mystified by the types."

For those who don't know what beer they might like (or need), I suspect that one of these days somebody will come up with an app to deal with the complexity. This would integrate data about your behavior and tastes, your health and medical data, your recent purchases, the season, time of day, and location ... and from all this data, and probably more, it would render a recommendation on the brew for you.

Then all you'd have to do is find out where to get it. Maybe the app could recommend that also ...



Re: Beer thirst
  • 1/12/2017 8:11:10 AM

@Ariella, that was along the lines of my point that styles of beer are good for those who know what they are drinking but for those that don't care to learn about methods for making beer tying brand names into the charts makes a lot of sense. Putting some brand names with the data would widen the audience greatly.

Now shifting into one of my loves (BBQ) use a darker beer for your brisket something like Shiner Bock which is getting more available outside Texas and is good for drinking after you've made your mop.  Also if you want more flavor use a thicker mop/baste and if you're BBQing or smoking wait until you're nearing the end of cooking before mopping.


Re: Beer thirst
  • 1/11/2017 1:01:14 PM

I'm still fairly mystified by the types. The first time I bought beer was in the fall, and that was for a brisket recipe. My family then decided that they don't like Heineken and tht it doesn't provide enough body for the meat. Since then, I've experimented a bit with different beers available for sale in single bottles usually for $1.50 at Trader Joe's. Some are described with some accuracy, but others just have these names that don't signify anything to those of us who are not familiar with these brews. 

Making data more usable
  • 1/11/2017 9:02:33 AM

I enjoy your posts and the ability to take data one step further, on this one though I might have done something like listed some of the most popular beers in each category in the title or on the larger view of the chart I would have pulled data from for example and listed it along side of the graph.  That way someone could say "I like X brand of beer" without trying to decipher if it's a Pale Lager or a Pilsner for example.  Most casual beer drinkers probably won't know what style they like but they'll be able to tell you what brands they like. 

Beer thirst
  • 1/11/2017 8:55:13 AM


Robert's blog post got my attention not just because it was about beer, but particularly used Pilsner as an example. Pilsners and lagers are my favorites types of beer.

I liked Robert's visualization redo, and the chart of beer categories with hyperlinks, although I can't say that I would need to have these to hand when I'm ordering up a pint in a pub or restaurant. But they're interesting to study ...

I kind of liked the amber color of the tiny graph boxes in the original graph. My only suggestion would be to use amber instead of red in the improved versions.

Anyway, this whole discussion makes me sorta thirsty ...


<<   <   Page 3 / 3