Monday, January 28, 2019

Wine consumption ranked by city

A few weeks ago I looked at wine consumption in different countries, pointing out that since 2010 the USA has been the world's biggest wine consumer. However, that situation is at least partly because the US is the third most populace country, so that even moderate consumption per person will quickly add up.

Another way that people have looked at wine consumption is based on individual cities. Once again, we would expect the largest cities to have the biggest consumption; but we can still ask the question as to the rank order of these large cities.

One organization that has been interested in trying to compile these data is the Wine and Spirit Institute of France’s INSEEC Business School. They have released their estimates for both 2016 and 2018; and these are shown in the graph. The estimates are in terms of millions of standard (750 ml) wine bottles.


Wine consumption per city estimated for 2016 and 2018

For each year, the only data we have are for the top 10 cities, and only eight of these appear on both lists — Naples and Madrid appeared only in the 2016 list, and Berlin and Tokyo only in the 2018 list.

Clearly, the estimates of consumption for 2018 exceed those for 2016, except for Paris (the pink line indicates equality). The biggest increases were for Milan and the conurbation of three cities from the Ruhr industrial region of north-west Germany (Essen, Dortmund and Duisburg now merge into each other). The data for Berlin also seem to show an increase, while the data for Naples and Madrid decreased.

Paris is the clear leader in terms of wine consumption, with second place not getting even close. Paris is not a particularly populous city compared to some of the others in the graph, but it has long been recognized as one of the most-visited cities in the world. Since tourists are likely to have a (French) wine or two while they are in town, trying to estimate per capita consumption would need to take that into account. Indeed, Decanter magazine did have a go at this without correcting for tourism, and their results clearly show the futility of trying to estimate how much wine each person in Paris (local + visitor) is drinking.

Mind you, wine is not cheap in Paris, no matter where you buy it. Indeed, French wine is probably cheaper in your homeland than it is in Paris. So, bring your money with you, if you want to indulge.


Finally, it is worth noting that the definition of a city is not straightforward (see Wikipedia), and I have no idea what version was used for the data collection here. However, no matter what definition we use, none of the cities listed above makes it into the top 10 largest cities in the world. Only Tokyo comes close (top 15), followed by London and New York (top 30). All of the other most-populous cities are not in countries noted for high per capita wine consumption.

Monday, January 21, 2019

Are words better than numbers for describing wine quality?

Last week, I wrote about using numbers to describe wine quality (The fundamental problem with wine scores). The advantage of expressing quality this way is that numbers conveniently have a simple linear order from minus infinity to plus infinity, so that they give the impression of being unambiguous. The disadvantage is that, in order to have a number mean anything worthwhile, it has to combine all of the various components of quality together, in which case its interpretation is actually ends up being quite ambiguous. Much of this discussion was recently summarized in: The demise in popularity of critical wine score pronouncements.


The only alternative we have is to use words. After all, we can use a whole set of words, describing each of the various quality components, so that we don't really need to combine multiple dimensions into a single word. As Kevin Brogle has noted: "I value the description more than the score. The score only tells you how much they liked it, not how much you will like it." So, this post is for Phllip White, who has always preferred words to numbers, and wine to either of them. Get well, mate.

Words versus numbers

Before we start, it is important to note that to the general wine-drinking population, words can apparently be just as effective as numbers. For example, consider this 2017 survey of purchase influences in the USA, from Wine Opinions. It suggests that a positive written wine review does just as well as 90+ score, in terms of influencing a purchase.


On the other hand, it is clear that the actual descriptions themselves are often not helpful. As an example, consider the poll conducted by One Poll for Lathwaites Wine 2013 survey: 66% of wine drinkers find wine descriptions UN-helpful in choosing wine. The people surveyed considered some words to be useful (eg. fresh, zesty, peachy) and some other words to not be (eg. brooding, haunting, tongue spanking).

There is a lesson here, I think, which is that words have serious limitations.

Some limitations of words

We regularly see blog posts from around the world noting that there are several different limitations to words when they are applied to wine (eg. Have we reached the end of wine criticism?). These include:
  • wine descriptions are not uniform, so we cannot compare them;
  • wine descriptions are too imprecise to help us evaluate wines;
  • wine descriptions are too flowery or pompous to be of practical value.
The last point may be a matter of taste, just like wine. For example, the above-mentioned Philip White has engendered this comment from Kim Brebach: "Philip White writes about wine in a way no one else ever has." This may very well be a good thing, which you can check out for yourself at his blog Drinkster.

However, the first two points can be addressed in a more objective manner. After all, we can still be quantitative about wine descriptions, even without numbers. For example, here is the Wine & Spirit Education Trust's recommended word "formula" for describing a wine.

WSET Systematic Approach to Tasting Wine

Not unexpectedly, most wine commentators go far beyond this. Indeed, they would probably have very few readers if they didn't show a bit more literary flair. And therein lies the basic issue of lack of uniformity and precision in wine descriptions.

Much of the information on this topic is summarized in the 2009 book by Adrienne Lehrer, Wine and Conversation, 2nd edition. The book discusses the large set of English words used for wine description (usually adjectives, and frequently metaphorical), and whether they each mean the same thing for different people. The author concludes that there is very little consensus about the meanings of wine words, although enologists are better able to agree on descriptions of wines than are lay people. However, most people are not able to describe the wines they have tasted so that other people could reliably match those descriptions to the actual wines.

Uniformity and precision

Steve Lay has noted: "Straw, tobacco, black cherries, currants, blackberries, crisp apples, floral, etc. These are just a few of the standard aroma definitions and some taste profiles of a wine; for which we can all agree." This is true, although the vocabulary actually goes way beyond these few words. Indeed, Lehrer's book starts with a list of 238 words "found in the wine literature". Therein lies the communication problem: how could we ever be either uniform or precise with so many words?

As a simple exercise, I have compared the 5-word descriptions from Journey Through Wine: an Atlas (Adrien Grant Smith Bianchi & Jules Gaubert-Turpin. 2017) and Wine Folly: the Essential Guide to Wine (Madeline Puckette & Justin Hammack. 2015). For each of the wines from 40 different grape varieties, I looked for words in common between the two books, interpreting this very loosely. The resulting counts are:
Words in
common
0
1
2
3
4
5
 Number
 of wines
6
17
11
3
3
0

So, most of the descriptions (70%) had only 1-2 words in common out of 5, and some had no words in common at all, which is pretty much what I was expecting. Even at this basic level, the words are neither uniform nor precise.


Solutions?

Sadly, I have no solutions, myself. Instead, I will present a couple of suggestions from other people.

Let's start with Oliver Styles: Wine writing's lack of judgment, who is apparently tired of repetitive tasting notes, and wants firm opinions rather than lists of adjectives:
while this ability to find and use new expressions for having a smell and a taste of something hints at a future on the Mills & Boon roster, we rarely get more understanding of whether the critic getting creative with language likes the wine or not. Basically, that's left to the number.
That is, instead of relying on florid word descriptions plus a number, we should actually write about whether we like the wine or not. Sounds simple, doesn't it? I doubt that it is, though.

Another alternative is to treat the data as multi-dimensional, just as I did last week for wine-quality scores. The bottom line is this: it still takes multiple words to describe all aspects of a wine's quality, and summarizing this in a word or two does not change anything. As I said for the situation of using numbers, we are still summarizing multiple dimensions (expressed as words, this time) into one dimension (a small set of words).

I pointed out last time that, under these circumstances, we actually need to draw graphs, in this case graphs of the word collection, rather than numbers. This idea has long been applied to wine descriptions, for example the early works by Louise S. Wu, R.E. Bargmann & John J. Powers (1977. Factor analysis applied to wine descriptors. Journal of Food Science 42: 944-952) and H. Heymann & A.C. Noble (1989. Comparison of canonical variate and principal component analyses of wine descriptive analysis data. Journal of Food Science 54: 1355-1358).

It is difficult seeing the wine-buying public going for this solution. However, I have shown one example picture below the line, along with a discussion, for anyone who might be interested.

Conclusion

It has been suggested that a wine-quality score is actually a word not a number (A wine rating is an adjective, not a calculation). If it is, then perhaps we should use actual adjectives, rather than supplying mathematical substitutes. Unfortunately, adjectives do not really seem to be up to the job of expressing quality, at least not in a uniform or precise manner. In which case, literary wine descriptions will continue to hold sway, with or without an attached score.

To quote Rusty Gaffney:
Some astute wine consumers look closely at the descriptors in a wine review and find it of paramount importance relative to the score. The description of the wine seems of greater usefulness, especially if the reader can see “between the lines”, and understands certain words or phrases that the writer uses that are a tip-off indicating a special wine. Wine descriptions would seem to be of most value to well-informed wine consumers, while others less knowledgeable about wine look more to scores for guidance.


Below is an ordination diagram, as an example of summarizing multi-dimensional wine information. It is taken from:
Alice Vilela, Bebiana Monteiro, Elisete Correia (2015) Sensory profile of port wines: categorical principal component analysis, an approach for sensory data treatment. Ciência Técnica Vitivinícola 30:1-8.
In this work, a panel of tasters evaluated each of each 28 port wines for a set of 23 characteristics. The characteristics could each be described by a word, representing colors, aromas, flavors, mouthfeel, balance and persistence.


Each of the 28 labeled points in the diagram represents one of the wines, labeled as either a White Port (BW), a Ruby Port (BR) or a Tawny Port (BT). The proximity of the points to each other represents how similar they were in wine style, based on their descriptions (by the evaluators). Note that the wines do cluster by type, with all of the whites grouped at the bottom-left, the rubies at the bottom-right, and the tawnies at the top of the diagram.

The 23 words represent the descriptions that were evaluated for each of the wines. The location of each word (or set of words) on the diagram tells us which wines they were associated with. For example, the description "Dried fruits" was associated mainly with Tawny Ports, whereas "Red fruits" was associated with Ruby Ports, and "Honey" was associated with White Ports. On the other hand, the descriptions "Woody" and "Golden" were associated equally with Tawny Ports and White Ports, but not with Ruby Ports; and "Red fruit flavor" was associated with Tawny Ports and Ruby Ports but not White Ports.

The diagram thus shows us a single picture of how all of the word descriptions related to all of the different wines.

Monday, January 14, 2019

The fundamental problem with wine scores

I have written several blog posts about wine-quality scores, pointing out that even though they are expressed as numbers they do not have many useful mathematical properties; and, to me, a score with no mathematical meaning is like trying to construct a Swedish sentence by knowing the words but not the grammar. However, what I have not done, until now, is point out the fundamental issue that leads to this situation in the first place. That is, I have previously pointed out effects, but not causes.

Before proceeding to discuss the cause, however, I will point out that many wine commentators seem to treat wine scores as nothing more than a convenient way to express their own personal preferences (ie. increasing score indicates increasing preference). Under these circumstances the scores have nothing to do with mathematics, at all. Preferences could just as easily be expressed with words; and in this case they probably should be. They certainly used to be, before the 1990s, and for some commentators they still are.


The basic issue

Put formally, wine scores represent multidimensional properties that have been summarized as a single point in one dimension.

Sounds good, doesn't it? Let's put it another way: the single wine-quality number is trying to do too many things all at once.

Whenever a critic tells us how they construct their scoring scheme, they usually list a series of characteristics of wines that purportedly contribute to quality (mainly based on color, aroma, palate and body). Formally, each of these characteristics is a "dimension" of any given wine's quality.

Here is an example, taken from Steve Charters and Simone Pettigrew (2007. The dimensions of wine quality. Food Quality and Preference 18: 997-1007).

The dimensions of wine quality

In terms of quality, most commentators are interested solely in the intrinsic dimensions. However, in order to describe a wine mathematically, we would need a number for each of these intrinsic dimensions. Given this collection of numbers, we would then have a complete description of any given wine's quality.

The situation

As a prime example, take the original UCDavis wine scoring system, which covers the score range 0-20.** The characteristics of quality and their associated numbers are:
Dimension
Appearance
Color
Aroma & bouquet
Volatile acidity
Total acidity
Sweetness
Body
Flavor
Bitterness
General quality
Score
2
2
4
2
2
1
1
2
2
2

There are 11 dimensions here, and we need all 11 numbers to completely describe any given wine's quality. That is, wine quality is multi-dimensional, and we need to "see" all of those dimensions in order to evaluate the wine.

However, rather than doing this, the UCDavis system summarizes the wine down to a single number — in this case, we add the numbers for each dimension, to get a score out of 20. That is, we reduce the multi-dimensional idea of quality down to a single point in one dimension — that dimension simply goes from 0 to 20, and the point on that dimension is the quality score.

The ensuing problem

The problem that arises from this situation actually applies any time we reduce a multi-dimensional concept down to a single dimension. I encountered this issue many times in my professional life as an environmental and evolutionary biologist,* so there is nothing unique about the situation as it arises in wine commentary.

The problem is this: many quite-different wines could end up with the same final score. Summarizing a set of numbers down to a single number must, by definition, lose most of the numerical information (the multiple dimensions become one dimension only). If a wine gets a score of 0, then we know the score for each dimension (it must be 0 in each case), and we have lost no information. The same applies for a wine that scores 20, as this must mean that the wine got the maximum score for each dimension. But for all other scores the situation is ambiguous.

Consider these two wines, which I have described using the 11 UCDavis dimensions listed above:
2 + 2 + 2 + 2 + 2 + 0 + 1 + 1 + 2 + 1 = 15
2 + 2 + 4 + 1 + 1 + 1 + 1 + 2 + 0 + 1 = 15
These would be two very different wines; but I would never know it from the final quality score.

So, you should now see why wine quality scores have a fundamental problem, if we try to treat them as mathematical concepts: how do we interpret the quality score? We have no way of knowing what the score represents in terms of the multi-dimensional concept of wine quality. Two identical scores could easily represent two very different wines.


A problem for all ratings systems

The problem discussed here is general. All ratings systems are one-dimensional, while the data on which they are based are multi-dimensional. A linear rating system makes no sense when you are combining different characteristics — we cannot combine multiple features into a single number in any way that makes much sense. That is, when we look at the final rating score we cannot tell which characteristics were important in producing it.

Take this simple situation, where value for money has two dimensions, quality and price:
A (high quality) a (expensive)
A (high quality) b (inexpensive)
B (low quality)  a (expensive)
B (low quality)  b (inexpensive)
How could I sensibly put these four groups in a single order based on value for money? We know which group is likely to be the best value for money, and we might put this at the top; and we know which is the worst value for money (Ab), and we might put this at the bottom (Ba); but what do we do with Aa and Bb in terms of value for money? If we did put them in some order, we would be doing so solely for the sake of doing so, not because it would be informative.

We have two totally different criteria, and combining them vitiates any attempt at a single order. The only system that would make sense would be multi-dimensional. That is, we should keep the ratings as Aa, Ab, Ba and Bb — the categories would this have meaning even though their order does not.

This is very similar to America's Got Talent, where the judges are trying to compare a magician with a pole dancer, and deciding which is "better". Better at what? Both of them are very good with their hands, but in very different ways! No wonder most of these shows worldwide end up being won by singers.

Wine shows

So, the issue for wine-quality ratings should now be clear. The ratings are based on trying to combine a series of different characteristics, some of which are very different from each other.

This explains why a wine can win a gold medal at one show and nothing at all at the next. The judges were combining the different quality dimensions in different ways, and thereby deciding which is best — that is all that the wine shows tell us.

The wine shows try to alleviate the problem a bit, by having a lot of different categories, based on all sorts of features (grape variety, wine style, vintage age, etc). This certainly helps, but it brings us back to the same problem of comparing two bottles of wine based on a series of vinous characteristics that are very hard to combine into a single number. And this approach certainly does not help at all with "best wine in show" awards.

A solution?

I have discussed multi-dimensional data previously in this blog. I pointed out at the time that, if we are going to take the numbers seriously, then we actually need to draw graphs of them, not reduce them to a single number:
Summarizing multi-dimensional wine data as graphs, Part 1: ordinations
Summarizing multi-dimensional wine data as graphs, Part 2: networks
It is difficult seeing the wine-buying public going for this solution, but I might discuss it in a future post.

An alternative solution?

It has sometimes been claimed that a wine score is not a number, but is more like an adjective. Well, it sure looks like a number to me, so this simply exacerbates the problem. If it is an adjective then it should be a word, not a number. I will discuss this in my next post, but as a preview: it still takes multiple words to describe all aspects of a wine's quality, and summarizing this in a word or two does not change anything — we are still summarizing multiple dimensions (expressed as words, this time) into one dimension (a small set of words).



* For example, in ecology Species Diversity is measured as a combination of two dimensions: (1) a count of the number of species, and (2) the abundance of each species. These two concepts are combined into a single number.

** Here is a more detailed overview of the UCDavis scoring scheme, taken from George Vierra (A better wine scorecard?).


Monday, January 7, 2019

The rise of the USA as the world's biggest wine consumer

Following last week's cautionary post (Is there truth in wine numbers?), we can now contemplate a few wine numbers about the USA and its position in the wine world.

Any given country's wine consumption is a product of the amount of wine each (adult) person consumes per year and the number of people in that country. To be No. 1, a country can either have a lot of people or they can each consume a lot of wine, or both. The USA is the third most populous country on the planet, after China and India, neither of which consumes a lot of wine per person (yet).

So, the idea that the USA is the No. 1 wine consumer is not unexpected. However, the question is when did it become No. 1? As the first graph shows, this event did not occur until very recently.

Top wine-consuming countries 1865-2014

The data are taken from Global Wine Markets, 1860 to 2016: a Statistical Compendium, compiled by Kym Anderson, Signe Nelgen and Vicente Pinilla. The graph covers the years 1865-2014 (horizontally), showing the estimated percentage of global wine consumption (vertically) for those countries that either currently account for >5% of the consumption or have accounted for >10% at some time in the past.

The graph makes it very clear that the patterns of wine consumption have changed dramatically over the past 150 years for many countries. This is not a result of changing population sizes (which have all grown), but instead reflects changes in per person (per capita) wine consumption. These changes are shown in the second graph.

Top wine-consuming countries per capita 1865-2014

Both France and Spain have shown a slow decline in consumption per person since the 1950s, although France also had particularly notable dips during both World War I and WWII. Italy's decline per person actually dates from the 1960s, although there were erratic changes in consumption during WWI. By contrast, Germany's rise in consumption dates from the end of WWII. The rise in per person consumption in the USA dates from the end of Prohibition, not unexpectedly. The only major wine-consuming country missing from the second graph is Portugal, which actually showed an increase in per person consumption until the end of WWII, followed by a slow decline starting in the mid 1970s.

So, the declining consumption in France and Italy combined with the rising consumption in the USA, finally resulting in the US taking the global lead in total consumption from the year 2010 onwards.

If you want to see some forecasts for the possible future of US wine consumption, they are discussed in 2019 U.S. alcohol consumption to increase while population growth stagnates.

Also shown on the second graph above (as the dashed line) is the consumption for Croatia, which had the globally highest per capita rate in 2014. Indeed, the per person consumption in Croatia has remained steady since the 1980s, unlike the western European countries discussed so far. Spanish per person consumption dropped below the current (Croatian) maximum way back in 1984, followed by the Italians in 2005, and the French and Portugese in 2013. The other two countries who currently match the Croatians, Portugese, French and Italians in their love of wine are the Moldovians and the Swiss.

Finally, it is worth illustrating just how far out in front the top three countries are, in terms of per person consumption. The final graph shows the per capita consumption (vertically) for the top 37 countries (ranked horizontally).


As you can see, all of the countries fit a nice simple (linear) mathematical model except the top three, where the people are still consuming far more wine per person than anywhere else.