Wine Project Report

Price vs. Quality

The following is a plot of price versus points of wines. The point system is the rating out of 100 that WineEnthusiast gave the wine. Please note that WineEnthusiast only posts ratings for wines rated 80 points or above. Also, I’ve restricted the x-axis because certain exorbitantly expensive wines made the rest of the relevant data difficult to see.

The below plot attempts to show how wine quality (judged by the WineEnthusiast point system) varies based on price.

ggplot(data = df, mapping = aes(x = price, y = points)) +
  geom_point(alpha = 0.05) + geom_smooth(color = "rosybrown1") + xlim(0, 1200)

As this graph shows, the point ratings do increase drastically with price from the $0 to $250 range, but as the price exceeds around $500, the trend plateaus – we aren’t seeing as much of a significant increase in quality. If you’re buying $500+ wine, it’s probably not so much about quality wine as it is proving to your friends that you’re extravagantly rich.

Ranking Average Point Scores by Country

The following is a ranked list of the average points per country that aims to show which countries produce the best wine. I also included a column showing the total number of wines reviewed, to ensure that any conclusions drawn are put into context.

avg_points_by_country <- df %>%
filter(!is.na(country)) %>%
group_by(country) %>%
summarize(mean = mean(points, na.rm=TRUE), total_wines = n())
kable(avg_points_by_country %>% arrange(desc(mean)))

Surprisingly (to me at least), England topped list, even higher than France and Italy. However, it’s worth noting that only 9 English wines were sampled. We can’t really assert that English wines are by and large the best wines because of their high point mean alone; we’ve only really looked at a very, very small sample of English wines (compared to, say, France, for which we’ve analyzed 21, 098 wines).

I decided that a more accurate way to visualize this would be a scatterplot, shown below. Please note that I’ve only used the 5 top-ranked countries, because there are too many countries to plot. I also overlaid a box plot so the means and distributions of data can be better understood.

top_5 <- df %>%
  filter(df$country == "England" | 
           df$country == "Austria" |
           df$country == "France" |
           df$country == "Germany" |
           df$country == "Italy") %>%
  mutate(country =  factor(country, 
                    levels = c("England","Austria","France","Germany","Italy"))) %>%
  arrange(country)
ggplot(data = top_5, mapping = aes(x = as.factor(country), y = points)) +
  geom_point(alpha = 0.05, position = "jitter") +
  labs(x = "Country", y = "Points") +
  geom_boxplot(color = "rosybrown1", alpha = 0)

Here we have a more accurate visual representation of the rank of average point values per country. We can see that while England has the highest mean, it has extremely few data points compared to the other countries ranking top 5.

Here’s a world map of the average point values per country. I’ve excluded Antarctica, because as far as I know, penguins aren’t very prolific winemakers.

world <- map_data("world") %>% left_join(avg_points_by_country, by = c("region" = "country"))
ggplot(data = world, 
               mapping = aes(x = long, y = lat, group = group)) +
  geom_polygon(aes(fill = mean)) +
  scale_fill_continuous(low="rosybrown1", high="darkred", 
                        na.value="snow2") +
  coord_map(xlim = c(-180,180), ylim = c(-60, 80)) +
  theme(
    axis.title.x = element_blank(),
    axis.text.x  = element_blank(),
    axis.ticks.x = element_blank(),
    axis.title.y = element_blank(),
    axis.text.y  = element_blank(),
    axis.ticks.y = element_blank(),
    panel.background = element_rect(fill = "white"))

Countries with a higher average point score are shown in a darker red (a nice Cabernet Sauvignon, perhaps?), while those with a lower average point score are shown in a paler pink. Countries that produced no wine in the dataset are shown in grey. Unsurprisingly, Europe seems to have the highest concentration of good-wine-producing countries. Visually, Canada, Australia, India, and South Africa are doing pretty good too. China’s doing its best, but its wines are apparently pretty bad (that said, we’ve only looked at 3).