It is of some interest to know what prices are paid for grapes, as this obviously affects wine prices. In particular, there are often big price differences between areas and between grape varieties. The problem, for me at least, is that this information is usually provided in a large table full of numbers, which is a bot overwhelming.
Even when a table is arranged in a visually helpful way (and many of them are not), I still find it awkward to find the patterns I am looking for. What I want is a picture, not a mass of numbers. This blog post provides such a picture, for the 2018 grape harvest prices in the USA.
The data come from Grape Connect, which describes itself as "a transparent and secure wine-grape, juice, and bulk wine marketplace". They have a neatly arranged table of the Average Grape Pricing by U.S. Appellation [2018 Harvest]. It "includes pricing data for the 25 most represented varietals in respect to online listing count from 1/01/18 until 9/26/18; it shows average per-ton pricing by AVA for the 2018 harvest."
The table is arranged in four columns: Grape variety, US state, American Viticultural Area (AVA), and $US price (weighted average Price-per-ton, with listing quantity as the weight). So, each row in the table presents the data for one grape variety in one AVA. This allows the reader to look up the price of a particular grape in a particular area, but it is very difficult to compare prices between areas and between varieties. The latter is also of interest.
In order to see pricing patterns among the grape varieties and AVAs, the table would be much better arranged with the varieties as the columns and the AVAs as the rows. Each cell in the table would then contain the price. That way, I could compare prices between grape varieties within each AVA by simply looking across a single row; and I could compare prices between AVAs for each grape variety by simply looking down a single column.
Even better would be a picture, not a table. Such a figure is called a Heat Map. This uses colors to represent the prices, rather than using the actual numbers. Here is an example, based on the Grape Connect data.
Click the image to see the full size (879x1600 pixels), where everything is readable.
The grape varieties are in alphabetical order, horizontally; and the AVAs are in alphabetical order within states, vertically. The prices (average weighted price per ton) are shown by the colors, as indicated by the scale in the bottom-right corner. The prices have been log10 transformed — the minimum price is $400 = 2.6 (orange); and the maximum is $9,500 = 4.0 (crimson). Missing combinations are colored white (ie. areas without price data for that grape variety).
The heat map quickly allows us to see how widespread is each grape variety, by how much of each column is filled — for example, Cabernet is more common in California than elsewhere, and Pinot is more common in Oregon*. We can also see which AVAs have lots of grape types, by how much of each row is filled — for example, the Washington AVAs tend to have more varieties than the other states (4 out of the 12 AVAs have at least 10 varieties**). We can also see the grape prices, by the colors — for example, Cabernet has the highest prices. These patterns are not necessarily unexpected, but the point is that we can easily see them using the heat map, which we cannot in the original table.
It might be even better to re-arrange the rows and columns of the heat map, to put the highest prices near each other. This can be done, but I do not currently have access to suitable software. The downside of doing this is that it would probably no longer allow the state prices to be seen (because AVAs from different states would be mixed).
* Cabernet sauvignon has prices for 60 of the AVAs in the table, which is 50% more than its nearest competitor (Syrah).
** Columbia Valley, WA and Rogue Valley, OR each have 15 grape varieties out of the 25 varieties that make it into the table.