Using data from 2019, dig into some visualizations around the world of wine ratings. Data Source: https://github.com/rfordatascience/tidytuesday/tree/master/data/2019/2019-05-28
| variable | class | description |
|---|---|---|
| country | character | Country of origin |
| description | character | Flavors and taste profile as written by reviewer |
| designation | character | The vineyard within the winery where the grapes that made the wine are from |
| points | double | The number of points WineEnthusiast rated the wine on a scale of 1-100 (though they say they only post reviews for wines that score >=80) |
| price | double | The cost for a bottle of the wine |
| province | character | The province or state that the wine is from |
| region_1 | character | The wine growing area in a province or state (ie Napa) |
| taster_name | character | The taster/reviewer |
| title | character | The title of the wine review, which often contains the vintage (year) |
| variety | character | Grape type |
| winery | character | The winery that made the wine |
Here we can use tidylo to look at the probabilities that individual words are used in a particular variety (Merlot, Chardonnay, etc), compared to the how often that word is used OUTSIDE of that variety. This gives us a list of language used to describe different wine types -- perfect if, like me, you have no idea what you're doing with wine other than enjoying it. Fake your way through any fancy party with this easy trick:
Looking at some simple scatterplots, we can look at how description length changes relative to the wine point score. Here, it looks like higher-scoring wines do tend to use more words! It looks like there's no real difference in the word length, so people generally aren't using more complex words for better-scoring wine.
What about the language that is used? If we assign categories to Low/Mid/High rated wines, what words are used? Let's look at the language used for price range, and for different point ratings on wines. Interesting -- the lower-rated wines seem to have almost sarcastic critical language associated with them; whereas lower-priced wines have more objective flavor descriptors.
These were filtered down to include only wines that were less than $50, as there are some really expensive single bottles that were throwing off my visuals. So -- limiting the data to the set of wines that I might actually be willing to buy someday.
Boxplots arranged by ratings and price show which wines consistently rate well -- noting countries like England and Canada have highly-rated, but expensive wines. Peru has lower-ratings, but also generally more inexpensive wines. What's interesting are the countries that have highly-rated, but inexpensive wines! If we look at the ratio of points to price, Ukraine, Romania, and India all have the most points-per-dollar, while England brings up the rear -- you really have to pay to get their highly-rated bottles!
descrip and model tbd
