Dataset:

• In this project, we analyze a medicine dataset with the goal of recommending fever medication based on patient characteristics. The dataset includes information on Temperature, Age, Gender, BMI, symptoms (Headache, Body Ache, Fatigue), lifestyle factors (Physical Activity, Diet), environmental measures (Humidity, AQI), etc.

Problem we are solving:

• Our goal is to clean and preprocess the data, perform thorough exploratory data analysis, build and evaluate a classification tree model to predict the recommended medication so that patients could easily decide which medicine they need to take.

We begin by loading the libraries needed for data import, manipulation, visualization, and model building.

# readr: For reading CSV files quickly
library(readr)

# dplyr: For efficient data manipulation (mutate, filter, summarize)
library(dplyr)

# ggplot2: For creating advanced visualizations
library(ggplot2)

# rpart: For building decision trees (classification and regression)
library(rpart)

# rpart.plot: For visualizing the decision trees built with rpart
library(rpart.plot)

# For kable(), used on our confusion matrix
library(knitr)       

# For kable_styling() used on our confusion matrix
library(kableExtra)  

Next, we import our dataset from a CSV file. We then display the first few rows to verify that the data loads correctly.

# Import the dataset from the specified file path, this is niche to my system
data <- read_csv(paste0("/Users/colepatola/Desktop/ASU-Year-3/",
                 "DAT-301/Project1/enhanced_fever_medicine_recommendation.csv"))

# Display the first six rows to confirm successful import
head(data) 

## # A tibble: 6 × 20
##   Temperature Fever_Severity   Age Gender   BMI Headache Body_Ache Fatigue
##         <dbl> <chr>          <dbl> <chr>  <dbl> <chr>    <chr>     <chr>  
## 1        36.1 Normal            89 Female  24   No       Yes       No     
## 2        37.5 Mild Fever        94 Male    26.6 No       No        No     
## 3        36.4 Normal            92 Male    27.8 No       No        No     
## 4        39.8 High Fever        66 Male    18.7 No       Yes       No     
## 5        39.3 High Fever        28 Male    21   No       Yes       No     
## 6        38.3 High Fever        48 Female  24.3 No       Yes       No     
## # ℹ 12 more variables: Chronic_Conditions <chr>, Allergies <chr>,
## #   Smoking_History <chr>, Alcohol_Consumption <chr>, Humidity <dbl>,
## #   AQI <dbl>, Physical_Activity <chr>, Diet_Type <chr>, Heart_Rate <dbl>,
## #   Blood_Pressure <chr>, Previous_Medication <chr>,
## #   Recommended_Medication <chr>

Before proceeding to data cleaning, we need a thorough understanding of our dataset’s structure and contents.

• str() function helps us see each column’s data type (numeric, character, factor).
• summary() function provides basic descriptive statistics (min, max, and median).
• This guides our decisions about which columns need factor conversion, whether there are potential outliers, and if any values are missing.

str(head(data) %>% select(1)) # Display the structure of the data frame to review data types and dimensions

## tibble [6 × 1] (S3: tbl_df/tbl/data.frame)
##  $ Temperature: num [1:6] 36.1 37.5 36.4 39.8 39.3 38.3

summary(head(data) %>% select(1))# Generate statistics to see the range, mean, median, etc

##   Temperature   
##  Min.   :36.10  
##  1st Qu.:36.67  
##  Median :37.90  
##  Mean   :37.90  
##  3rd Qu.:39.05  
##  Max.   :39.80

# Convert relevant character columns to factors with specific levels where applicable
data <- data %>% 
  mutate(
    Fever_Severity = factor(Fever_Severity, levels = c("Normal", "Mild Fever", "High Fever")),# Convert to a factor
    Gender = factor(Gender), # Convert 'Gender' to a factor
    Headache = factor(Headache, levels = c("No", "Yes")),         # Convert  to a factor
    Body_Ache = factor(Body_Ache, levels = c("No", "Yes")),       # Convert to a factor
    Fatigue = factor(Fatigue, levels = c("No", "Yes")),           # Convert to a factor
    Chronic_Conditions = factor(Chronic_Conditions, levels = c("No", "Yes")),    # Convert to a factor
    Allergies = factor(Allergies, levels = c("No", "Yes")),                      # Convert to a factor
    Smoking_History = factor(Smoking_History, levels = c("No", "Yes")),          # Convert to a factor
    Alcohol_Consumption = factor(Alcohol_Consumption, levels = c("No", "Yes")),  # Convert to a factor
    
    # 'Physical_Activity' is an ordered factor
    Physical_Activity = factor(Physical_Activity, levels = c("Sedentary", "Moderate", "Active")), 
    
    Diet_Type = factor(Diet_Type),  # Convert 'Diet_Type' to a factor without order
    
    Blood_Pressure = factor(Blood_Pressure, levels = c("Low", "Normal", "High")),  # Convert to an ordered factor
    
    Previous_Medication = factor(Previous_Medication),  # Convert to a factor, same as diet type, no order
    Recommended_Medication = factor(Recommended_Medication) # Convert to a factor (target variable), no order
  )

# Check for any missing values in each column; this will return a named vector with counts of NA values
missing_counts <- sapply(data, function(x) sum(is.na(x)))
missing_counts  # Print out the missing value counts for review (nothing was missing)

##            Temperature         Fever_Severity                    Age 
##                      0                      0                      0 
##                 Gender                    BMI               Headache 
##                      0                      0                      0 
##              Body_Ache                Fatigue     Chronic_Conditions 
##                      0                      0                      0 
##              Allergies        Smoking_History    Alcohol_Consumption 
##                      0                      0                      0 
##               Humidity                    AQI      Physical_Activity 
##                      0                      0                      0 
##              Diet_Type             Heart_Rate         Blood_Pressure 
##                      0                      0                      0 
##    Previous_Medication Recommended_Medication 
##                      0                      0