Visual Analytics Final Project: Portuguese Wine and Which Components Correlate to Higher Quality

 

Visual Analytics Final Project

Tyler House

2023-04-27

library(ggplot2)
library(ggcorrplot)
library(ggmap)
library(ggraph)
library(ggpubr)
library(tidyverse)
library(ggthemes)
library(dplyr)
library(readr)
library(tinytex)
library(latexpdf)
library(stats)

winedata <- read_csv("C:\\Users\\tyler\\Desktop\\Spring 2023\\Visual Analytics\\Final Project\\winequality-red.csv")

str(winedata)

## spc_tbl_ [1,599 × 12] (S3: spec_tbl_df/tbl_df/tbl/data.frame)
##  $ fixed acidity       : num [1:1599] 7.4 7.8 7.8 11.2 7.4 7.4 7.9 7.3 7.8 7.5 ...
##  $ volatile acidity    : num [1:1599] 0.7 0.88 0.76 0.28 0.7 0.66 0.6 0.65 0.58 0.5 ...
##  $ citric acid         : num [1:1599] 0 0 0.04 0.56 0 0 0.06 0 0.02 0.36 ...
##  $ residual sugar      : num [1:1599] 1.9 2.6 2.3 1.9 1.9 1.8 1.6 1.2 2 6.1 ...
##  $ chlorides           : num [1:1599] 0.076 0.098 0.092 0.075 0.076 0.075 0.069 0.065 0.073 0.071 ...
##  $ free sulfur dioxide : num [1:1599] 11 25 15 17 11 13 15 15 9 17 ...
##  $ total sulfur dioxide: num [1:1599] 34 67 54 60 34 40 59 21 18 102 ...
##  $ density             : num [1:1599] 0.998 0.997 0.997 0.998 0.998 ...
##  $ pH                  : num [1:1599] 3.51 3.2 3.26 3.16 3.51 3.51 3.3 3.39 3.36 3.35 ...
##  $ sulphates           : num [1:1599] 0.56 0.68 0.65 0.58 0.56 0.56 0.46 0.47 0.57 0.8 ...
##  $ alcohol             : num [1:1599] 9.4 9.8 9.8 9.8 9.4 9.4 9.4 10 9.5 10.5 ...
##  $ quality             : num [1:1599] 5 5 5 6 5 5 5 7 7 5 ...
##  - attr(*, "spec")=
##   .. cols(
##   ..   `fixed acidity` = col_double(),
##   ..   `volatile acidity` = col_double(),
##   ..   `citric acid` = col_double(),
##   ..   `residual sugar` = col_double(),
##   ..   chlorides = col_double(),
##   ..   `free sulfur dioxide` = col_double(),
##   ..   `total sulfur dioxide` = col_double(),
##   ..   density = col_double(),
##   ..   pH = col_double(),
##   ..   sulphates = col_double(),
##   ..   alcohol = col_double(),
##   ..   quality = col_double()
##   .. )
##  - attr(*, "problems")=<externalptr>

summary(winedata)

##  fixed acidity   volatile acidity  citric acid    residual sugar 
##  Min.   : 4.60   Min.   :0.1200   Min.   :0.000   Min.   : 0.900 
##  1st Qu.: 7.10   1st Qu.:0.3900   1st Qu.:0.090   1st Qu.: 1.900 
##  Median : 7.90   Median :0.5200   Median :0.260   Median : 2.200 
##  Mean   : 8.32   Mean   :0.5278   Mean   :0.271   Mean   : 2.539 
##  3rd Qu.: 9.20   3rd Qu.:0.6400   3rd Qu.:0.420   3rd Qu.: 2.600 
##  Max.   :15.90   Max.   :1.5800   Max.   :1.000   Max.   :15.500 
##    chlorides       free sulfur dioxide total sulfur dioxide    density     
##  Min.   :0.01200   Min.   : 1.00       Min.   :  6.00       Min.   :0.9901 
##  1st Qu.:0.07000   1st Qu.: 7.00       1st Qu.: 22.00       1st Qu.:0.9956 
##  Median :0.07900   Median :14.00       Median : 38.00       Median :0.9968 
##  Mean   :0.08747   Mean   :15.87       Mean   : 46.47       Mean   :0.9967 
##  3rd Qu.:0.09000   3rd Qu.:21.00       3rd Qu.: 62.00       3rd Qu.:0.9978 
##  Max.   :0.61100   Max.   :72.00       Max.   :289.00       Max.   :1.0037 
##        pH          sulphates         alcohol         quality    
##  Min.   :2.740   Min.   :0.3300   Min.   : 8.40   Min.   :3.000 
##  1st Qu.:3.210   1st Qu.:0.5500   1st Qu.: 9.50   1st Qu.:5.000  
##  Median :3.310   Median :0.6200   Median :10.20   Median :6.000 
##  Mean   :3.311   Mean   :0.6581   Mean   :10.42   Mean   :5.636 
##  3rd Qu.:3.400   3rd Qu.:0.7300   3rd Qu.:11.10   3rd Qu.:6.000 
##  Max.   :4.010   Max.   :2.0000   Max.   :14.90   Max.   :8.000

head(winedata, n = 4)

## # A tibble: 4 × 12
##   `fixed acidity` `volatile acidity` `citric acid` `residual sugar` chlorides
##             <dbl>              <dbl>         <dbl>            <dbl>     <dbl>
## 1             7.4               0.7           0                 1.9     0.076
## 2             7.8               0.88          0                 2.6     0.098
## 3             7.8               0.76          0.04              2.3     0.092
## 4            11.2               0.28          0.56              1.9     0.075
## # ℹ 7 more variables: `free sulfur dioxide` <dbl>,
## #   `total sulfur dioxide` <dbl>, density <dbl>, pH <dbl>, sulphates <dbl>,
## #   alcohol <dbl>, quality <dbl>

#For this numerical analysis we will run correlation tests between all variables and the quality index column which is on a scale of 3-8 from this dataset. From the results we will select the three variables with the greatest correlation under the Pearson statistical measure, and we will visualize these variables in Part Two.

#Correlation Analysis
##Fixed Acidity and Quality Index

fac <- cor(winedata$`fixed acidity`,winedata$quality, method = "pearson", use = "complete.obs")

##Volatile Acidity and Quality Index

vac <- cor(winedata$`volatile acidity`,winedata$quality, method = "pearson", use = "complete.obs")

##Citric Acid Concentration and Quality Index

cac <- cor(winedata$`citric acid`,winedata$quality, method = "pearson", use = "complete.obs")

##Residual Sugar and Quality Index

ras <- cor(winedata$`residual sugar`,winedata$quality, method = "pearson", use = "complete.obs")

##Chloride Concentration and Quality Index

cc <- cor(winedata$chlorides,winedata$quality, method = "pearson", use = "complete.obs")

##Free Sulfur Dioxide Concentration and Quality Index

fsd <- cor(winedata$`free sulfur dioxide`,winedata$quality, method = "pearson", use = "complete.obs")

##Total Sulfur Dioxide Concentration and Quality Index

tsd <- cor(winedata$`total sulfur dioxide`,winedata$quality, method = "pearson", use = "complete.obs")

##Density and Quality Index

dd <- cor(winedata$density,winedata$quality, method = "pearson", use = "complete.obs")

##pH and Quality Index

phc <- cor(winedata$pH,winedata$quality, method = "pearson", use = "complete.obs")

##Sulphate Concentration and Quality Index

scc <- cor(winedata$sulphates,winedata$quality, method = "pearson", use = "complete.obs")

##Alcohol Concentration and Quality Index

ac <- cor(winedata$alcohol,winedata$quality, method = "pearson", use = "complete.obs")

##Now we will compile these correlations into a table

correlationdata <- c("ac","cac","cc","dd","fac","fsd","phc","ras","scc","tsd","vac")
r <- c(ac,cac,cc,dd,fac,fsd,phc,ras,scc,tsd,vac)

correlationtable <- data.frame(correlationdata,r)

colnames(correlationtable) <- c("Compositional Element", "Pearson Coefficient (r)")

#Correlation Bar Plot Source

pearsonbarplot <- ggplot(correlationtable, aes(x = correlationtable$`Compositional Element`, y = correlationtable$`Pearson Coefficient`)) +
  geom_bar(stat = "identity", fill = "#009d86", color = "black") +
  labs(y = "Pearson Coefficient (r)", x = "Compositional Element of the Wine") +
  theme_minimal()

pearsonbarplot

#The results of the Pearson Correlation Analysis show that the Citric Acid Concentration, Sulphates Concentration and Alcohol Content are the most positively correlated with Quality.

#Using the p-values now, we are going to make a correlellogram so to speak, of the different variables against eachother. With the use of hierarchical clustering I can view the independant relationships of each individual variable against one another and in reference to the data found previously I can infer which combinations of elements might contribute to lower quality

correlation <- round(cor(winedata),1)

c.matrix <- cor_pmat(winedata)

lowercorrelation <- ggcorrplot(correlation, hc.order = TRUE, lab= TRUE, type = "lower", method = "circle",colors = c("#6D9EC1", "white", "#E46726"))

highercorrelation <- ggcorrplot(correlation, hc.order = TRUE, lab= TRUE, type = "upper")


##Lower Half Correlation

lowercorrelation

##Upper Half Correlation

highercorrelation

#Statistical Results:
##The elements with the highest correlation to quality of the wine from this data set are the Citric Acid, Sulfate and Alcohol Concentrations. Without having to ask, I know why higher alcohol makes a better wine, who doesn't like a little more buzz per serving right? As for the other two components, I found in some research outside the data set that sulfates and sulfites (because of the presence of Sulfur Dioxide) in wine are the preservation agent and the enhancer of flavor in a multitude of ways. It was said from a few articles written from winemakers, that these elements are found as a byproduct of yeast fermentation which is a process that makes any number of ethyl alcohol derivatives. Chemically, the presence of greater amounts of sulfur based byproducts is related to a higher alcohol production and therefore, a higher quality on average.

##The next element was Citric Acid and these days, people are getting testier and testier, with their personalities and mannerisms, especially with their drinks of choice. Citric Acid is a compound that is notably bitter when introduced in higher amounts to any drinks composition but is used mainly to prevent ferric hazes within the wines and also to produce more flavor to flatter, sweeter wines that are more common of the Iberian Peninsula. It adds a sharpness to the flavor and also allows for microbial agents to metabolize and ferment at a higher rate which produces more of the sulfates and alcohol.

##The final element, and the element that is most heavily correlated with the previous two would be the concentration of alcohol within the wine. The Sulfur byproducts and citric acid concentrations relative to one another and the other compounds will assist in driving the alcohol gravity higher, as free ethyl alcohols are free to bind with sulfurs and other free compounds, as well as the citric acid being able to allow microbes to metabolize and produce more alcohol quicker.

#In my time in Spain and Portugal, many of the wine I had was of a bittersweet and strong profile, with a heavy alcohol content and a sour aftertaste. I am sure now that the experience I had was in part to the data taken from wines of the same regional profile that I experienced while on vacation

#My final say, is the sommelier was correct, simple acid levels, sulfur byproduct levels and alcohol concentrations being the highest correlated with wine qualities. But, from my experience I want to clarify that the wines that I tasted that were tastier, happened to be higher in alcohol. As the alcohol content rose, the body of the wine covered more flavor profiles, which backs up my research stating that the other compounds tend to be related to alcohol production quantities and rates.

#Overall the initial hypothesis stands to be partially correct from this data based on the variables they share, such as acidity and alcohol content, however the data added to what the sommelier said, by showing a high correlation of sulfites and sulfates to the other variables and to overall quality. Sugar being one of his talking points, when compared to the data in this analysis, showed an almost non existent correlation with p < 0.02. Regardless, his information was valid to an extent and therefore leaves more questions to what relationships of compounds are optimal for a truly fine wine.

##Short Summary
#This relates to the work in this class because it is always possible to find correlations in any data and make inferences on that data, if that it is cleaned enough and broad enough to cover multiple variables and above all else, is able to show a wide array of population members in the data set that make for a solid analysis, I wanted to have fun with an interest of mine, by finding real world data to work with and test against a fond experience of my past, I took that trip 6 years ago and I am still very fond of the times I have spent there, and I have been back since. It relates to the methods of this course because essentially, the point of a visualization is to see the trends and relationships within data rather than just a table. It allows for us to possibly see undiscovered trends to test numerically that can help enrich our understanding of the data we might be working with. This was the underlying fundamertals to this analysis and that brings it to the end, thank you for a wonderful semester Dr. Friedman !!